TW201249496A - Image collating apparatus, patient positioning apparatus, and image collating method - Google Patents

Abstract

An objective of this invention is to realize a high-accuracy two-stage pattern matching (two-stage collating) even in a case that the number of tomographic images of a three-dimensional present image is fewer than that of a three-dimensional reference image when positioning a patient in a radiation therapy. This invention includes a collation processing part (22) which collates a three-dimensional reference image (31) and a three-dimensional present image (36) and calculates a body position correction amount so that the position/posture of an affected part in the present image (36) may be corrected to match the position/posture of the affected part in the reference image (31). The collation processing part (22) has a primary collating part (16) for performing a primary collation which collates the reference image (31) with the present image (36); and a secondary collating part (17) for performing a secondary collation which collates a specified template area (40) generated from either one of the reference image (31) or the present image (36) based on the result of the primary collation, with a specified retrieval object area (42) generated from the other one of the reference image (31) or the present image (36), which is different from the generation origin of the specified template area (40), based on the result of the primary collation.

Description

201249496 6. EMBODIMENT OF THE INVENTION The present invention relates to a radiotherapy apparatus that performs cancer treatment by irradiating radiation of x-rays, 7-beams, and the like to an affected part of a patient, and uses ct ( Computerized tomography) An image collation device for imaging data; and a patient positioning device that uses a sharp image collation device to position the patient to radiation that illuminates the radiation. [Prior Art] In recent years, in the field of radiation therapy devices for cancer treatment, cancer cell therapy devices using particle beams such as protons and heavy ions have been developed and constructed (here, they are called particles). The beam therapy device is well known. Compared with the conventional radiation therapy using x-rays, r-rays, etc., the particle beam treatment of the beam treatment can concentrate the irradiation on the cancerous part, that is, it can match the opening of the affected part. The beam of the pinpoint illuminates the particle beam, and it can be treated without slamming the cells. In particle beam therapy, it is important to irradiate the particle beam with high precision to the affected part such as cancer. Therefore, it is used in particle beam therapy. The fixation device or the like fixes the patient so as not to be displaced from the treatment table of the treatment room (irradiation room). In order to accurately position the affected part such as cancer in the radiation irradiation range, it is first used to the laser index (laser) The setting of the rough fixation of the patient such as p〇inter) and the like, and then the precise positioning of the affected part of the patient is performed using an X-ray imaging or the like. Patent Document 1 proposes Species: reference is not in the portrait (x-ray fluoroscopic image day) and the present day image (X-ray receiver and using the captured image of the day)

The portraits of any of S 4 323646 201249496 all specify the same number of identical positions of the plurality of montages, but perform a two-stage pattern-matching to generate a positioning for driving the treatment table. Information bed positioning • Devices and their positioning methods. The primary pattern matching sets a second setting area approximately the same size as the second setting area for the two-dimensional current image (where the first setting area includes isocenter (beam irradiation) set for the two-dimensional reference image The area of the center), then sequentially moving the second set area in the area of the two-dimensional present image, and comparing the two-dimensional reference image in the first set area to the second set area in each position of the second set area The two-dimensional current drawing # inside is inserted into the second setting area of the two-dimensional current image having the most similar to the two-dimensional reference image of the second region. The quadratic pattern matching is to compare the -dimensional current image in the second set region extracted in the -type pattern matching with the two-dimensional reference image in the first setting region of the month ij The two images are the most consistent way to match the pattern. (Patent Document 1) (Patent Document 1) Japanese Patent No. 3748433 (0007 to 0009, paragraph 0049, Fig. 8, and Fig. 9) [Summary of the Invention] Since the shape of the affected part is a three-dimensional three-dimensional shape, the three-dimensional imaging can achieve a higher accuracy than the case of using the two-dimensional imaging in terms of positioning the affected part at the position of the affected part. In general, X-ray CT (c〇mputed 5 323646 201249496) will be used when the treatment plan is completed.

Tomography) The image is used to determine the three-dimensional shape of the affected part. In recent years, it has been desired to provide an X-ray CT apparatus in a treatment room to use an X-ray CT image obtained by an X-ray CT apparatus and an X-ray CT image of a treatment image at the time of treatment. Requirements for positioning. This is because the X-ray translucent image towel 'this f is the soft-faced part of the affected part usually does not appear very clear, the peak is compared with the system (4) relative position to age, and vice versa, using x-ray CT image Therefore, it is not necessary to directly perform the alignment between the affected parts captured in the X-ray CT imaging. Therefore, it is considered to expand the reference image and the current artifact into a three-dimensional artifact in the conventional two-stage pattern matching. The three-dimensional reference image and the three-dimensional image now contain a plurality of tomographic images (slice images) taken by a x-ray CT apparatus. Because there are X-ray exposures, etc., it is false; 3D now has fewer artifacts, so when comparing, you must have a more detailed image of the third image. A comparison of three-dimensional current portraits with key image information that is abbreviated to the three-dimensional reference image. The segment pattern matching 'can compare the two-dimensional reference image with the same density of the image information and the two-dimensional current image, but the three-dimensional reference image and the three-dimensional current image with different imaging information (four degrees) Fortunately, the simple transfer of the imagery dimension (d-) of the technique from two dimensions to one dimension has the problem of not achieving a two-stage pattern matching. Change: "There is no such thing as it is: in the past, simply the three-dimensional reference image from the set m-th field (10) is used for the three-dimensional current image in the second setting area, and is simply extracted. The second setting 323646 6 201249496 The three-dimensional current image portion in the region is compared with the three-dimensional reference image in the first set region to match the pattern in which the two images are most consistent. Therefore, the object of the present invention is to achieve high-precision two-stage pattern matching even when the number of tomographic images of the three-dimensional current artifact is smaller than that of the three-dimensional reference image when the patient is subjected to radiotherapy ( Two-stage check). (Means for Solving the Problem) The image collation apparatus of the present invention includes a three-dimensional imaging input unit, which is a three-dimensional reference image taken during a treatment plan for radiation therapy, and is taken at the time of treatment. The three-dimensional current image is read in separately; and the check processing unit checks the three-dimensional reference image and the three-dimensional current image, and calculates that the position and posture of the affected part in the three-dimensional current image are consistent with the position and posture of the affected part in the three-dimensional reference image. The amount of body position correction. The check processing unit has a primary check unit that performs primary pattern matching from the three-dimensional reference image with respect to the three-dimensional current image, and a secondary check portion that performs the matching from the three-dimensional reference based on the result of the primary pattern matching. The predetermined template area generated by one side of the image or the three-dimensional current image is compared with the other one of the three-dimensional reference image or the three-dimensional current image which is different from the basis of the generation of the predetermined template region according to the result of the primary pattern matching. A quadratic pattern matching of the generated predetermined search target area. (Effect of the Invention) The image checking device of the present invention first performs a pattern matching from the three-dimensional reference image matching to the three-dimensional current image, and then generates a predetermined template region and a predetermined one based on the result of the one-type matching. The search target area, s 7 323646 201249496, performs the quadratic pattern matching between the search target area and the template area, so that even if the number of tomographic images of the three-dimensional current artifact is smaller than that of the three-dimensional reference image, high precision can be realized. Two-stage pattern matching. [Embodiment] Embodiment 1 FIG. 1 is a view showing a configuration of an imaging collation apparatus and a patient positioning apparatus according to Embodiment 1 of the present invention. Fig. 2 is a view showing the overall configuration of a machine relating to the imaging apparatus and the patient positioning apparatus of the present invention. In Fig. 2, 1 shows a CT simulation room for performing a treatment plan performed before radiotherapy, in which there is a CT gantry 2 and a top plate 3 of a CT imaging photographic bed, and the patient 4 is lying horizontally. On top of the top plate 3, CT image data containing the treatment plan for the affected part 5 is taken. On the other hand, 6 denotes a treatment room for performing radiation therapy, in which there is a CT pylon 7 and a rotating treatment table 8, the upper portion of the rotary treatment table 8 has a top plate 9, and the patient 10 is placed on the top plate 9, and The CT imaging data for the positioning of the affected part 11 at the time of treatment is taken. Here, the positioning means calculating the position of the patient 10 and the affected part 11 at the time of treatment from the CT image data for the treatment, and calculating the position correction for the position of the patient 10 and the affected part 11 in accordance with the treatment plan. The amount is then carried out in such a manner that the affected part 11 at the time of treatment comes to the radiation irradiation center 12 of the radiation therapy. The alignment is achieved by driving control of the rotary treatment table 8 in a state where the patient 10 is lying on the top plate 9 to move the position of the top plate 9. The rotary treatment table 8 can perform a translation correction of six degrees of freedom for translation and rotation, and the top plate 9 of the rotary treatment table 8 can be rotated by 180 * 8 323646 201249496 degrees 'to the top of the rotary treatment table 8 (-) ( In Fig. 2, the solid line indicates that one of the treatment positions of the irradiation head 13 is shifted (in the second, "radiation 2" diagram, although the money is not the CT imaging position and the treatment position) 4 (10) The case of the relative positional relationship of degrees, but it is not limited to this configuration. The positional relationship between the two is also different. The position of the CT image and the angle of the treatment are used. The CT image used for the position is transmitted to the positioning computer 14. The CT used for the treatment is used as a dimensional reference image, and the image is used as a three-dimensional image verification device 29 and patient positioning in the present invention. Set: Elephant. The computer software related person existing in the computer 14 checks the body position correction amount (translation amount, rotation amount), and the patient device 30 includes not only the image checking device 29 but also Further, according to the posture correction amount (4), the function of the parameters of the respective drive shafts of the wire control treatment table 8 (hereinafter simply referred to as the treatment table 8) is calculated. The person positioning device 30 controls the treatment table 8 according to the comparison made by the image checking device 29: thereby guiding the target portion of the particle beam treatment to the beam irradiation of the treatment device Center 12. The positioning done in traditional radiotherapy is based on the DRR (Digitally Reconstructed) generated from the CT image data from the treatment plan.

Radiography) The positional shift 1 is calculated as a fluoroscopy fluoroscopy image taken at the same time as the DRR squint and a fluoroscopy fluoroscopy image taken at the treatment room during treatment. In the fluoroscopy perspective, the affected part of the soft tissue usually does not appear to be clear, so basically the relative position with the head of the bone is used to perform the alignment. The positioning of the ct room 6 image data has the following two?: The CT pylon 7 is set in use, and before the treatment = 匕 because it is being treated... The alignment is performed on the CT image data, and the image is poor. The treatment plan is done in the opposite position of the affected part. "The straight-hearted part can be explained, and the position correction device 29 and the patient-constituting image collation device and the patient positioning σ " in the positioning device 30 of the present embodiment can be explained. Figure 1 shows the system. _, the image collation device 29; ^== the three-dimensional portrait input unit 21 between the analytic units, the collation processing unit 22, the bedding material read 23, and the collation result rounding unit 乂 the result display unit 3上上:疗_一二计昼基::::: In the treatment of 昼 Τ Τ Τ Τ Τ Τ ' ' ' ' ' ' ' ' ' : : : : : : : : : : : : : : : : ' : ' ' ' ' ' ' ' ' ' Treatment * Hand and wheel (four), the position of the use of the Bayer eight features: from the point of view of inhibition of ray exposure, the number of complex fault images (also known as sliced images) is less. The invention is configured to first perform a pattern matching from the three-dimensional reference key matching to the three-dimensional current image, and then generate a predetermined template area and a predetermined search target area according to the result of the one-type matching, and then use the predetermined The model area and the quadratic pattern in the same direction or in the opposite direction Pattern matching the configuration phase. Phase two pattern matching can be described by the pair

S 323646 10 201249496 The matching parameter at the time of pattern matching is different from the comparison parameter in the case of quadratic pattern matching, and the processing speed and accuracy are achieved. For example, there is a pattern matching for a wider range with a coarser resolution, and then using the found template area or the search object area to perform quadratic pattern matching on the reduced range with finer resolution. The method. Next, the three-dimensional imaging input unit 21 will be described. The three-dimensional imaging input unit 21' is a D1COM (Digital Imaging and Communications) image group that is captured by an X-ray CT apparatus and configured as a plurality of tomographic images. In Medicine) The image data of the form (slice group) is read as a three-dimensional volume data. The CT image data used for treatment planning is the three-dimensional volume data for the treatment of the measurement, that is, the three-dimensional reference image. The CT image data for positioning is the three-dimensional volume data at the time of treatment, that is, the three-dimensional current image. Moreover, the CT image data is not limited to the dic〇M form, and may be in other forms. The collation processing unit 22 collates the three-dimensional reference image and the three-dimensional current image in the pattern matching (the pattern matching) to calculate the posture correction amount that matches the position and posture of the affected part in the three-dimensional current imaging with the position and posture of the affected part in the three-dimensional reference image. The collation result display unit 23 causes the collation processing unit 22 to collate the result (the body posture correction amount to be described later and the two-dimensional current image that has been moved by the posture correction amount are superimposed and displayed on the three-dimensional reference image). Like, etc.) is not visible on the monitor surface of the computer 14. The collation result output unit 24 is a correction amount obtained when the collation processing unit 22 rides the collimation of the three-line quasi-image and the three-dimensional current image, that is, the amount of correction of the body position calculated by the collation processing unit 22 (translation amount, (four) amount ) Call out. ^Therapy (4) Parameter calculation 323646 11 201249496 The output 26 will check the output value of the bran output unit 24 (translation three axes [△ X, △ Υ, ΔΖ], rotation = ~ axis [ΔΑ, ΔΒ, △〇 In total, six degrees of freedom are converted into parameters for controlling the axes of the treatments 4 ..^ y 13 8 , that is, the gray μ mountains such as v, which are used to control the axes of the osmotic treatment 〇8. The treatment table 8 drives the drive means of each axis of the treatment table 8 based on the 乂 λ / 〇 treatment table control parameters calculated by the treatment table control parameter. In this way, the position correction amount Α, Λ, which allows the position of the affected part to match the treatment plan, can be calculated, and then the target portion 11 during the treatment is irradiated to the radiation treatment of the radiation therapy, and the alignment is performed. The collation processing unit 22 includes a position and posture converting unit 25, a primary collating unit 16, a secondary collating unit & β^7, and a reference template region generating unit 18. The position and posture changing unit 25 records the position and orientation of the object data by $1 in the case of the ~^ pattern matching or the quadratic pattern matching. The primary check unit 16 performs a one-dimensional matching of the three-dimensional reference image with respect to the three-dimensional current left A. / 疋 in the portrait. The secondary collating portion 17' is a predetermined template region generated from the three-dimensional reference image or the third image, and the result is matched with the result according to the -sub pattern matching (four) and the predetermined template according to the result of the -type pattern matching. The base of the generation of the different dimensions of the current search object generated by the other side of the Q domain 'in the second two: the undertype match. The check processing unit 22 will be described in detail using the first to ninth drawings.帛3目系_Lin invention invention 1 + three-dimensional reference - 昼 image and reference vertical image. Fig. 4 is a view showing a three-dimensional present portrait in the first embodiment of the present invention. Fig. 5 is a view for explaining a primary sample matching method in the first embodiment of the present invention. Fig. 6 is a diagram showing the relationship between the reference image area in the pattern matching method of Fig. 5 and the cut image of the image of r· Λ 323646 12 201249496. Fig. 7 is a view showing a single extraction region of a sliced image extracted by the primary pattern matching method in the first embodiment of the present invention. Fig. 8 is a view for explaining a quadratic pattern matching method in the embodiment of the present invention. Fig. 9 is a diagram for explaining the relationship between the reference image template area and the slice artifact in the quadratic pattern matching method of Fig. 8. The reference template area generating unit 18 of the collation processing unit 22 generates the reference image template area 3 from the three-dimensional reference image 31 by using the shape of the affected part (affected part information) which is input at the time of the treatment. The three-dimensional beauty of the Zhuixiang 31 is composed of a plurality of sliced images 32. In the third drawing, an example in which the two-dimensional reference image 31 is composed of five sliced images 32a 32b 32c 32d and 32e is shown in order not to obscure the description. The shape of the affected part is entered as a 闭f Interest: attention area 35 in a form in which the sliced image is a closed contour surrounding the affected part. The area including the aforementioned closed contour may be set to, for example, an circumscribed square 34, and the cube area including each circumscribed square 34 may be set as a template area. This template area is used as a reference template area 33. The primary collating unit 16 of the collation processing unit 22 performs primary pattern matching in which the reference image template region 33 is matched to the three-dimensional current image 36. The two-dimensional present image 36 shown in Fig. 4 is an example of three sliced images 37a, 37b, and 37c. The current image area % shown in Fig. 5 is expressed as a cube containing a slice of the image 37a, 37b, 37c. As shown in Fig. 5, in the current imaging area 38, the reference image template area 33 (33a, 33b, 33C) is moved in a raster scan manner, and is calculated and compared with the three-dimensional current image 1 3 6夕| Relevant value. Any correlation value in the image comparison (image check) can be used in terms of correlation values. S. 13 323646 201249496 The reference image template area 33a is moved in a progressive scan manner along the line sweeping path 39a along the sliced Chad, and on the 1st 37a. Similarly, the Ukrainian review has a commentary field 33b on the sliced image 37b along the sweeping path ^ outside the image of the sample area, the reference image area 33c is t μ ^, the image 37c The edge scan path 39c moves in a progressive scan. The dragon is rotated by the α ° r ' sweeping path 39b 39c in order to simplify the display without complicating the drawing. When the --type matching is performed, as shown in Fig. 6, each of the slice images 53 constituting the reference image template area 33 is imaged with the slice image 37 constituting the current imaging area 38, respectively. Check. The sliced image 53 is an image cut out by the reference frame 2 in the slice image 32 of the three-dimensional reference image 31. The reference image template area 33 is composed of five sliced images 53a, 53b, 53c, 53d, 53e corresponding to the five sliced images 32a, 32b, 32c, 32d, 32e in the three-dimensional reference image. Therefore, at the time of the pattern matching, the slice images 37a' of the three-dimensional current image 36 are respectively subjected to the five slice images 53a, 53b, 53c, 53d, 53e of the reference image template region 33. The portrait is checked. Further, the image collation is performed in the same manner as the slice images 37b and 37c of the three-dimensional current image 36. The primary collating unit 16 extracts the primary extraction region 43 from each of the slice images 37 of the three-dimensional current image 36 by one pattern matching (this primary extraction region 43 includes the current imaging region 38 and the reference image template region 33). The area with the highest correlation value). As shown in Fig. 7, the extracted area 43.a is extracted once from the sliced image 37a of the three-dimensional now image 36. And the extraction regions 43b, 43c are extracted once from the sliced images 37b, 37c of the two-dimensional current image 36. Then, it is generated that the one-time drawing area 43a, 43b, 43c is included in 14 323646 201249496, one of which is drawn in the keying area 42 as the search target area to be used in the quadratic pattern matching. In this way, the primary collating unit 16 generates a one-time drawing of the search target area to be used in the quadratic pattern matching in the imaging area 42. However, in the state before the positioning, the three-dimensional reference image 31 does not coincide with the posture of the three-dimensional current image 36 (rotation three axes), so the number of slices in the three-dimensional current image 36 as in the simple progressive scan of FIG. In a rare case, it is impossible to perform a high-precision matching in which the angular offset is detected, but there is no problem in extracting the one-time extraction area 43 required for the secondary pattern matching. Therefore, in the one-shot matching, the angular offset is not detected and the correlation value is calculated, and the subsequent quadratic pattern matching is performed with the high-precision matching detected by the angular offset. Next, the secondary pattern matching will be explained. In the quadratic pattern matching, the position and posture conversion unit 25 of the collation processing unit 22 generates a position and posture conversion template obtained by converting the position and orientation of the reference image template area 33 generated from the three-dimensional reference image 31. Area 40. As shown in Figs. 8 and 9, the quadratic pattern matching system adds the posture change amount (rotation three axes) of the reference image template area 33 as a parameter. The second collation unit 17 is formed by the position and posture conversion template region 40 obtained by the position and posture conversion unit 25, and the three-dimensional current image 36 having a small number of sliced images is present in the imaging region 42. Between, the high-precision matching that also includes the angular offset factor is included. This design allows for a high-precision two-stage pattern that also includes angular offset factors. By using the narrower range of 15 323646 201249496 containing the region found in the primary pattern matching as the object of the quadratic pattern matching search range, it is possible to use a coarser resolution and A wide range of one-time extraction area 43 obtained as a target for one type matching is present in the image area 42 to perform a second-level match with a finer resolution, and the pattern matching can be shortened. Time required. The sputum sample shown in Fig. 8 appears in the sculpt area 42, which is a cube of three primary extraction regions 43a, 43b, 43c. In the present ready-made package, the position of the reference image template region after the transformation is changed, and the region 4〇a is moved in a progressive scan manner by changing the scanning path 39a in the secondary extraction region of the sliced image 37a. Similarly, the position and orientation change template area 4卟 on the 43a is converted, and the first extraction area 43b of the posture image 37b is swept in the slice along the scan path, and the position and posture are converted. The bit-shaped area 40c is moved in a progressive scan manner in the one-time extraction area of the sliced image 37c. Among them, the upper portion along the 39c system simplifies the display so as not to complicate the drawing. Tian Suo 39b, when the quadratic pattern is matched, as shown in Fig. 9, the section 41 of the position and posture change template area 40 and the section image of the current image area of the manipulating section are shown by Fig. 9 One extraction area of 37 - under-extracted line image check. In addition, it can also be used in sectioning image 55 (using one; between the image area 42 in the three-dimensional image of the current image of the cut moon, you appear on the ι image 37 and cut out the image), and the section 41 Perform an image check. The section 41' of the 昼 field 40 which is centered by the position and posture is generated from the number of slices of the three-dimensional reference image 3i and the number of slices of the template area. For example, the data of Section 41 is a two-dimensional reference image of the image.

S 323646 16 201249496 31 of the plurality of slices are like 32 cutters. In general, the data density of the section 41 of the position and orientation conversion template area 40 is different from the data density of the one extraction area 43 of the three-dimensional current image 36, but the correlation value may be calculated for each element of the section 41. Further, the section 41 of the position/posture conversion template area 40 may be a data supplemented in such a manner as to increase the data density of the first extraction area 43 of the three-dimensional current image 36. Here, the two-stage pattern matching method of the first embodiment will be described. First, the reference template area generating unit 18 of the collation processing unit 22 generates the reference image template area 33 from the three-dimensional reference image 31 (reference image template area generating step). Then, the primary collating section 16 performs a pattern matching of the reference thumbnail image area 33 with respect to the three-dimensional current image 36 (primary pattern matching step). The primary pattern matching system constitutes each of the slice images 53 of the reference image template area 33, and collates with the image of the slice image 37 constituting the current imaging area 38. The primary collating unit 16 scans the reference image template area 33, and calculates the correlation value between the current imaging area 38 and the reference image template area 33 for each scanning position (correlation value calculation step), since one pattern matching The extraction extraction area 43 is extracted by extracting the correlation value between the current image area 38 and the reference image template area 33 so as to include the highest area (the extraction area extraction step). Then, the primary collating unit 16 includes a single extraction region 43 in such a manner that each of the slice images 37 constituting the current imaging region 38 is generated, and one of the search target regions to be used in the quadratic pattern matching is generated. Keying area 42 (search object generation step). The two-stage pattern matching method of Embodiment 1 is an inclusion base.

S 17 323646 201249496 The quasi-昼 model area produces a step of the second plastic matching step of the step-by-step pattern. The one-time matching step is a packet closing value calculation step, a single extraction region extraction step, and a retrieval object generation step. Then, the secondary collating unit 17 of the collation processing unit 22 converts the position and posture conversion template region 40 in which the position and orientation of the reference image template region 33 is converted from the position and posture conversion unit 25, with respect to the three-dimensional current image %. Sub-extraction occurs in the imaging area 42 to perform a quadratic pattern matching (two (four)-like four-matching step). The quadratic pattern matching is to generate a plurality of sections 41 (section generation steps) of the position and orientation transformation template area 40 that have been transformed, and to form a section of the image area 42 in each section 41'. The primary extraction area 43 or the slice image 55 of the image 37 is subjected to an image collation with the section 41. The secondary collating unit 17 scans the position and orientation conversion template area 40, and calculates the correlation value of the plurality of sections 41 that appear in the imaging area 42 and the position and posture transformation template area 4〇 for each scanning position (correlation value). calculation steps). Then, the position/posture conversion unit 25 converts the position and posture into a position and posture different from the previous position and posture (positional ampoule conversion step), and the second collation unit 17 generates a plurality of positions and postures in the position and posture conversion template area 40. Section 41 (section generation step)' and scanning the position and orientation transformation template area 4〇, and calculating a plurality of breaks appearing in the imaging area 42 and the position and posture transformation template area 40 for each scanning position. The correlation value of the face 41 (correlation value calculation step). The secondary collating unit π of the collation processing unit 22 selects the correlation value as the highest three-dimensional reference image of the calculated correlation value and the positional orientation of the three-dimensional current image 323646 18 201249496 potential relationship (position and posture information) as the optimal solution. (Best deselection step). In this way, it is possible to match the three-dimensional reference image with the three-dimensional image of the three-dimensional current image. The quadratic pattern matching step includes a section generation step, a correlation value calculation step, a position and orientation transformation step, and an optimal solution selection step. After the pattern matching is completed, the collation processing unit 22 calculates that the three-dimensional reference image 31 and the three-dimensional current image 36 are collated from the position and posture of the positional posture change template region 40 which is the highest among the calculated correlation values. Position correction amount (translation amount, rotation amount) at the time of the body position (step correction calculation step). Then, the collation result display unit 23 displays the body posture correction amount and the key image displayed by superimposing the three-dimensional current image on which the body posture correction amount has been superimposed on the three-dimensional reference image on the monitor surface of the computer 14. Further, the collation result output unit 24 outputs the posture correction amount (translation amount, rotation amount) when the collation processing unit 22 collates the three-dimensional reference pupil image 31 with the three-dimensional current imaging image 36 (body position correction amount output step) . Then, the treatment table control parameter calculation unit 26 converts the output value (translation three axes [ΔX, ΔΥ, ΔΖ], and the rotation three axes [ΔΑ, ΔΒ, AC] total six degrees of freedom) of the verification result output unit 24 into The parameters of the respective axes of the treatment table 8 are controlled, that is, the parameters for controlling the axes of the treatment table 8 (the treatment table control parameter calculation step) are calculated. Then, the treatment table 8 drives the drive means of each axis of the treatment table 8 in accordance with the treatment table control parameters calculated by the treatment table control parameter calculation unit 26 (the treatment table driving step). Because the image matching device 29 of the first embodiment performs the pattern matching of the three-dimensional reference image 31 with respect to the three-dimensional current image 36, then 19 323646 201249496 is based on the result of the primary pattern matching, from the three-dimensional reference image 31. A template area for predetermined quadratic pattern matching, that is, a position and orientation transformation template area 40 is generated, and a three-dimensional current image 36 is generated to include the primary extraction area 43 therein for use in the secondary pattern matching. The predetermined search target area, that is, the one-time extraction appears in the imaging area 42, so that even if the number of tomographic images (the number of sliced images) of the three-dimensional current image 36 is smaller than that of the three-dimensional reference image 31, high precision can be achieved. The two-stage type match. According to the image collation apparatus 29 of the first embodiment, even in the case where the number of tomographic images (the number of sliced images) of the three-dimensional current image 36 is smaller than that of the three-dimensional reference image 31, high-precision two-stage pattern matching can be realized. Therefore, the number of tomographic images of the three-dimensional current image 36 taken by the X-ray CT apparatus at the time of alignment can be reduced, and the amount of X-ray exposure received by the X-ray CT apparatus for the patient can be reduced. Further, the anamorphic collation apparatus 29 of the first embodiment generates a priming in the anamorphic region 42 based on the result of performing the pattern matching of the three-dimensional reference anamorphic image 31 with respect to the three-dimensional squall image 36, and This one-time drawing, which is narrower than the current imaging area 38, appears in the imaging area 42 as a retrieval object, so that a coarser resolution can be used to perform a pattern matching for a wider range of objects, and then the use includes matching in one pattern. One extraction of the primary extraction region 43 found in the imaging region 42 occurs to perform quadratic pattern matching with a finer resolution, and the time required for pattern matching can be shortened. According to the patient positioning device 30 of the first embodiment, the position and posture of the patient and the positional posture of the treatment device 20 323646 201249496 can be set to the position and posture of the patient according to the posture correction amount calculated by the imaging verification device 29. The positional posture at the time of the spot treatment plan is such that the image portion η at the time of treatment can be brought to the beam irradiation center 12 of the radiation therapy. Further, in the patient positioning device 3 of the first embodiment, the potential transforming unit Μ is matched to the image pattern area 33 obtained from the three-dimensional reference image 31 to the number of tomographic images (the number of slice images). Base = Portrait 31 Less 3D Now Painting | 36 is a very suitable position position ^Change the plate area 40, and the two-stage type (4) that also shifts the angle can be realized. The accuracy of the material is based on the two-dimensional reference image 36 of the embodiment i, which is obtained by the three-dimensional reference picture taken during the treatment of the radiation therapy. Read the gold 仏 Department 21 separately; check the third secret reading 31 / use the human two-dimensional image input to let the three-dimensional book now -:; quasi; two: two curry Φ "position and position in the three-dimensional reference image ^ (four) The position and posture—the body posture correction amount verification processing unit U′ and the verification processing unit 22 includes: performing a three-dimensional reference image μ with respect to the two-dimensional current image 36; The predetermined plate area (position and posture transformation template area 40) generated from the three-dimensional reference image 31 or the three-dimensional current image 36 is formed according to the result of the one-time pattern matching. The predetermined pattern area (bit 1 posture change template area 4_ generates a second type of matching of the predetermined search target area 42 generated by the two-dimensional reference 4 image 31 or the three-dimensional current image 36) The collating portion 17' thus even the tomographic image of the two-dimensional now-image 36 Number ratio 3D benchmark 21

S 323646 201249496 It is also possible to achieve a two-stage type matching of South precision like the 31/month. According to the patient positioning device 3 of the first embodiment, the image collation device 29 and the posture correction amount calculated by the imaging collation device 29 are used to calculate the parameters for controlling the parameters of the respective axes of the treatment table 8. The image control parameter calculation unit 26 includes the three-dimensional reference image 31 taken during the treatment of the radiation therapy and the two-dimensional current image 36 taken during the treatment. The three-dimensional image input unit 21 collates the two-dimensional reference image 31 and the three-dimensional current image 36, and calculates the posture correction amount that causes the position and posture of the affected part in the three-dimensional current image 36 to coincide with the position and posture of the affected part in the three-dimensional reference image 31. The processing unit 22 is checked. The collation processing unit 22 has a primary collation unit 16 that performs one-time matching of the three-dimensional reference image 31 with respect to the three-dimensional current image 36; and performs a three-dimensional reference image 31 or three-dimensionally based on the result of the one-time pattern matching. The predetermined template area (position and posture change template area 40) generated by one of the 36 sides is different from the basis of the generation of the predetermined template area (position and posture change template area 4) with respect to the result of the one-type pattern matching. The secondary reference image of the predetermined search target region 42 generated by the other of the dimensional reference image 31 or the two-dimensional image 36 is matched to the secondary collation portion 17, so that even the two-dimensional 昼胄36 (four) layer 昼 image ratio When the three-dimensional reference image 31 is small, high-precision alignment can be performed. ~ 诼 对 对 万 万 万 万 万 万 ... ... ... ... ... ... ... 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩 钩A pattern matching step comprising a three-dimensional reference image 31 relative to a first phase of the two-dimensional current image 36 ε 323646 22 201249496; and performing a pattern from the one-dimensional matching from the three-dimensional reference The predetermined template area (position and posture transformation template area 40) generated by one of the 31 or three-dimensional current portraits 36 is generated from the predetermined template area (position and posture conversion template area 40) with respect to the result of the primary pattern matching. The quadratic pattern matching step of the quadratic pattern matching image of the predetermined search target region 42 generated by the other of the three-dimensional reference image 31 or the three-dimensional current image 36, so that even the three-dimensional present image 36 has the number of tomographic images Compared with the three-dimensional reference image 31, high-precision two-stage pattern matching can be realized. Embodiment 2 In the two-stage pattern matching of Embodiment 2, the pattern matching from the three-dimensional reference image 31 to the three-dimensional current image 36 is performed first, and then the image is obtained from the three-dimensional image according to the result of the one-type matching. The generated template area for matching the secondary pattern matching (now the template area 44) is converted to the posture and the reference image area 47 obtained by converting the position and posture of the three-dimensional reference image 31 as a search target. The stencil-like template area 44 is matched to the quadratic pattern of the posture-converted reference anamorphic region 47. The quadratic pattern matching is a pattern matching in the opposite direction to the one-type matching. Fig. 10 is a view for explaining a primary pattern matching method in the second embodiment of the present invention, and Fig. 11 is for explaining a reference image area and a slice image in the primary pattern matching method of Fig. 10. Diagram of the relationship. In the primary pattern matching according to the second embodiment, the primary collating unit 16 performs the search including the rotation three axes to obtain the posture change amount. The present portrait area 38 shown in Fig. 10 is represented as a cube containing three cuts 23 323646 201249496 • sheet portraits 37a, 37b, 37c. The position and orientation conversion template areas 40a, 40b, and 40c of the reference image template area in the second embodiment are the areas in which the position and posture are converted by the position and posture conversion unit 25. The initial position and posture is the default state, for example, the parameters of the three axes of rotation are all in the state of 〇. The position and posture conversion template area 4〇a obtained by changing the position and orientation of the reference image template area is moved in the same manner as the scanning path 3 along the scanning path 3, and the position is made The position and orientation conversion template area 40b which has been subjected to the potential transformation is moved in a progressive scanning manner along the scanning path 39b on the sliced image 37b, and the position and posture conversion template area 40c which has been transformed by the position and posture is sliced. The image is moved in a progressive scan along the scan path* on the image 37c. Among them, the scanning paths 39b and 39c are simplified in order not to complicate the drawing. - while changing the position and posture - while performing the three-dimensional current image 36, the images 37a, 37b, 37c are correlated with the position and posture transformation template area. For example, the respective axes of the three axes of rotation are changed by a predetermined amount of change or rate of change', and then the correlation is calculated, and then moved to the next scanning position, and then the correlation calculation is performed. As shown in Fig. u, the secondary collating unit 16 performs an image collation between the cross-section 41 of the position and posture changing template region 40 and the slice image 37 constituting the current imaging region %. The cross-section 41' of the position-and-posture conversion template area 4' is a section which is formed by cutting the position and posture conversion template area on a plane parallel to the blade-shaped image 32 of the three-dimensional reference image of the first position and posture. The generator is generated from a plurality of slice images 32 of the three-dimensional reference image 31 (section generation step). The section 41 can be generated by, for example, the method described in the description of the method of 24 323646 201249496. That is, the data of the section 41 can be set to be cut out from the plurality of slice images 32 constituting the three-dimensional reference image 31. Further, the section 41 of the position and posture conversion template area 40 can be included to allow The data density is increased in the same way as the data density of the three-dimensional image 36. Then, the primary collating section 16 produces the present portrait template area 44 to be used in the quadratic pattern matching. The primary collating unit 16 obtains the cross-section 41 of the position-posture conversion template region 40 having the highest correlation value, for example, from the results of the search including the rotation three axes for each of the slice images 37a, 37b, and 37c. The positional posture changes the posture change amount of the template region 40 and the extraction region of the sliced image 37 corresponding to the cross section 41. The primary check unit 16 generates the current image template area 44 including the extracted area of the three-dimensional current image having the highest correlation value from among the extracted extraction areas of the respective sliced images. The jaw-like panel area 44 is now a two-dimensional image. Then, as shown in FIG. 12, the position and posture converting unit 25 of the collation processing unit 22 converts the posture of the entire three-dimensional reference image 31 by the posture change amount obtained when the current image template area 44 is generated. The three-dimensional posture transformation reference image 45 after the posture conversion is generated, that is, the posture transformation reference imaging region 47 is generated. Fig. 12 is a view showing a three-dimensional reference image after the posture conversion in the second embodiment of the present invention. The sliced images 46a, 46b, 46c, 46d, and 46e are sliced images in which the postures of the slice images 32a, 32b, 32c, 32d, and 32e are changed by the posture change amount, respectively. Then, as shown in Fig. 13, the secondary collating unit 17 changes the reference image 45 of the three-dimensional posture after the posture change of the present image sample area 44,

S 25 323646 201249496 That is, the amperometric conversion reference imaging area 47 t is scanned and scanned in a progressively scanning manner along the scanning path 49, so that only the translational shift can be detected at high speed. Fig. 13 is a view for explaining a quadratic pattern in the second embodiment of the present invention. The posture-converted reference image area -47 which is transformed by the posture is expressed as a cube including five sliced images 46a, 46b, 46c, 46d, 46e. The check execution face 48 corresponds to a posture (this posture is a posture in which the pose corresponding to the slice image of the three-dimensional current image 36 has the southernmost correlation value). The image plane, that is, the posture in the posture transformation reference image region 47 and the posture corresponding to the slice image 37 of the three-dimensional current artifact % have the same posture. The secondary collating unit 17 generates a predetermined collation execution surface 48 (check execution surface generation step) in the posture conversion reference imaging region 47 from the plurality of slice images 46 of the three-dimensional posture conversion reference image 45. The collation execution surface 48 can be generated by, for example, the method described in the first embodiment. That is, the data of the verification execution surface is set to be a plurality of sliced image cutouts constituting the three-dimensional posture transformation reference image 45. In addition, the verification execution surface 48 may also be a data included in a manner that allows the data density to be the same as that of the current image area 44. The two-stage pattern matching method of Embodiment 2 will be summarized. First, the position and posture conversion unit 25 of the collation processing unit 22 generates a position and posture conversion template area 40 (position and posture conversion template area generation step) obtained by performing position and orientation conversion from the three-dimensional reference image 31. Then, the primary collation unit μ of the core processing unit 22 performs a pattern matching of the position and orientation conversion template area 4 相对 with respect to the three-dimensional current image 36 (primary pattern matching step 323646 26 201249496). The sub-pattern matching is performed for each of the ten images 37 constituting the current imaging area 38, each time the position and posture change template region milk is set to change (every time the position and posture transformation step is performed), Position: The cross-section 41 of the posture changing template region 40 (the step of generating the cross-section) is subjected to the collation of the cross-section between the cross-section 41 of the posture-converting template region 40 and the sliced image 37 constituting the current imaging region. Further, the primary collating unit 16 calculates the correlation value appearing in the imaging region publication and the positional k-changing template region 40 each time the position and posture of the position-and-posture conversion template region 40 is changed (correlation value calculation step). And, the primary collating unit 16 calculates the correlation value appearing in the imaging region 38 and the position and posture transformation template region 4〇 each time the position and posture transformation template region 4〇 is scanned, and the primary pattern matching is used to generate the correlation value. The position and posture of the current image area 38 and the position and posture transformation template area 4A are the highest, and the extraction area of the template area 40 is included in the present image template area 44 (now the template image area generation step). Then, the collation processing unit 22 converts the posture of the entire three-dimensional reference image 31 by the position and posture change amount obtained when the current image template area 44 is generated by the position and posture conversion unit 25, and generates a three-dimensional posture transformation after the posture conversion. The reference image 45, that is, the posture transformation reference image region 47 (posture transformation reference image region generation step) is generated, and then the secondary collation portion 17 executes the current image template region 44 twice with respect to the posture transformation reference image region 47. Pattern matching (secondary pattern matching step). The quadratic pattern matching generates a collation execution surface 48 in the collation execution surface generation step, and then performs the collation execution surfaces 48 and 27 generated in the collation execution surface generation step. 323646 201249496 Now the image of the template area 44 is checked. At the time of this image check, the collation execution surface 48 and the current template area 44 are calculated while the current template area 44 is being translated (not rotated). Correlation value (correlation value calculation step in the quadratic pattern matching 'reconciliation processing unit 22's secondary collation unit Π The correlation value is the calculated correlation value The positional relationship (position and posture information) of the highest three-dimensional posture transformation reference image 45 and the current jaw image region 44 is selected as the optimal solution (optimal deselection step), and thus can be realized by two-stage matching. The three-dimensional reference image 31 is matched with the two-dimensional image of the three-dimensional current image 36. The two-stage pattern matching method of the second embodiment includes the position and orientation transformation template region generation step, the sub-pattern. The matching step, the posture transformation reference image area generation step, and the second miscellaneous peaks.

The positive amount, and the second amount of the body position correction amount, the positive amount (the amount of translation, the amount of rotation) (the body of the result display portion 23 makes the two-dimensional present portrait of the posture corrected to the second

The check processing unit 22 is placed on the three-dimensional surface. Moreover, the verification result output unit 28

S 323646 201249496 The position correction amount (translation amount, rotation amount) when the reference image 31 is collated with the three-dimensional current image 36 is output (the body position correction amount output step). Then, the treatment table control parameter calculation unit 26 converts the output values of the collation result output unit 24 (translation three axes [ΔΧ, ΔΥ, ΔΖ], and the rotation three axes [ΔΑ, ΔΒ, Δ (total total six degrees of freedom) into The parameters for controlling the axes of the treatment table 8, that is, the parameters for controlling the axes of the treatment table 8 (the treatment table parameter calculation step) are calculated. Then, the treatment table 8 is calculated based on the treatment table control parameter calculation unit 26. The treatment table control parameter drives the driving device of each axis of the treatment table 8 (the treatment table driving step). Because the imaging matching device 29 of the second embodiment performs the position and orientation conversion template region 40 of the three-dimensional reference imaging 31 with respect to The three-dimensional image 36 now also includes a pattern matching of the image collation including the rotation three axes, and then generates a template area for the quadratic pattern matching from the three-dimensional current image 36 according to the result of the one-time matching. In other words, even if the number of tomographic images (the number of sliced images) of the three-dimensional image 36 is smaller than that of the three-dimensional reference image 31, the high-precision two-stage pattern can be realized. Further, in the image matching device 29 of the second embodiment, the three-dimensional posture conversion reference image 45 (three-dimensional reference image after the posture conversion) is generated from the three-dimensional reference image 31, that is, the posture conversion reference is generated. The image area 47 can be directly matched by using the two-dimensional current template area 44 so as to be parallel to the posture-shifting reference image area 47 without rotational movement. In the pattern matching, only the correlation value of each parallel movement is calculated, so that the speed of the second pear-like matching can be achieved as compared with the case of calculating the correlation value of each rotation movement and parallel movement of 29 323646 201249496. The third embodiment is based on the use of a human body database (atlas model) to generate the reference image template region 33 for the primary pattern matching in the first embodiment or the basis of the position and orientation conversion template region 40 in the second embodiment. The image template area 33 is different from the first and second embodiments. Fig. 14 is a view showing the configuration of the image collation apparatus and the patient positioning apparatus according to the third embodiment of the present invention. The image collation apparatus 29 of the third embodiment differs from the image collation apparatus 29 of the first and second embodiments in that the human body database input unit 50 and the average template area generating unit 5 are in the third embodiment. The positioning device 30 includes an image collation device 29 and a treatment table control parameter calculation unit 26. The human body database input unit 50 acquires an atlas model from a memory device such as a database device, and an average template region generating unit 51. The average template region 54 is cut out from the organ portion of the human body database corresponding to the affected parts 5, 11 of the patient 4, 10. The reference template region generating portion 18 of the collation processing portion 22 is by the average template region 54 is matched with the three-dimensional reference book image 31 to automatically generate the reference image area of the reference image area. Performing the separation using the above-described reference image template area 33 is performed! The two-stage pattern matching or the two-stage pattern matching in the second embodiment. In this way, even if the information on the affected part (the shape of the affected part, etc.) is provided on the three-dimensional reference image, the two-stage pattern matching can be realized.

S 30 323646 201249496 In addition, it is also conceivable that the average template region generating portion 51 cuts out the average template region of the human body from the organ portion of the human body database corresponding to the affected portions 5, 11 of the string 4'10. In the case of the two-dimensional average template area 54, Cosway: a two-dimensional average template area is subtracted, and then a plurality of two-dimensional output-plate areas are output to the verification processing unit 22. The US standard template area generating unit 18 of the collation processing unit 22 automatically generates the base: plate area 33 by pattern matching the plurality of two-dimensional flat template=fields with the two-dimensional reference image 31. Decent [simplified description]

Fig. 1 is a view showing the configuration of an image collating device positioning device according to an embodiment i of the present invention. W Fig. 2 is a view showing the overall configuration of the apparatus related to the imaging collation apparatus and the patient positioning apparatus of the present invention. Fig. 3 is a view showing a three-dimensional reference image and a reference image template area in the first embodiment of the present invention. Fig. 4 is a view showing a three-dimensional current image in the first embodiment of the present invention. Fig. 5 is a view for explaining the primary pattern matching method in the first embodiment of the present invention. Fig. 6 is a view for explaining the relationship between the reference image area and the slice image in the primary pattern matching method of Fig. 5. Fig. 7 is a view showing a single extraction region of a sliced image extracted by the secondary pattern matching method in the first embodiment of the present invention. Fig. 8 is a view for explaining a method of arranging a secondary type sample 31 323646 201249496 in the first embodiment of the present invention. Figure 9 is a diagram showing the relationship between the basis of the model and the sliced image of the secondary pattern of the Japanese (4) 8 figure. Fig. 10 is a view for explaining the primary pattern matching method in the second embodiment of the present invention. Fig. 11 is a view for explaining the relationship between the reference image template area and the slice artifact in the one-shot matching method of Fig. 10. Fig. 12 is a view showing a three-dimensional reference image after the posture conversion in the second embodiment of the present invention. Fig. 13 is a view for explaining a quadratic pattern matching method in the second embodiment of the present invention. Fig. 14 is a view showing the configuration of an imaging collation apparatus and a patient positioning apparatus according to a third embodiment of the present invention. [Main component symbol description] 1 CT simulation room 2,7 CT tube holder 3,9 top plate 4, patient 5, 11 affected part 6 treatment room 8 rotating treatment table 12 beam irradiation center 13 irradiation head 14 computer

S 32 323646 201249496 16 Primary check unit 17 Secondary check unit 18 Reference template area generation unit 21 Three-dimensional image input unit 22 Check processing unit 23 Check result display unit 24 Check result output unit 25 Position and posture change unit 26 Treatment table control parameter calculation Part 29 Image reconciliation device 30 Patient positioning device 31 Three-dimensional reference image 32a to 32e Sliced image 33, 33a, 33b, 33c Reference image area 34 External quadrilateral 35 ROI (Note area) 36 Three-dimensional now image 37a~37c Sliced image 38 now image area 39a~39c Scan path 40, 40a, 40b, 40c Position and orientation change template area 41 Section 42 Once drawn in the image area 43 Once extracted area 33 323646 201249496 44 Now the image area 45 Pose change reference image. 46a to 46e Slice image: 47 Pose change reference image area 48 Check execution surface 49 Scan path frustration 50 Human body database input unit 51 Average template area generation unit 53 Slice image 55 Slice image 34 323646

Claims (1)

201249496 VII. Patent application scope: 1. An image checking device, which has: - a three-dimensional image input unit, which measures the treatment of radiation therapy; the three-dimensional reference image taken during the treatment and the treatment The obtained three-dimensional current image is read separately; and the * check processing unit checks the three-dimensional reference image and the three-dimensional current image, and calculates the position and posture of the affected part in the three-dimensional current imaging and the three-dimensional reference image. The collation processing unit includes: a primary collating unit that performs primary pattern matching from the three-dimensional reference image with respect to the three-dimensional current image; and a secondary collation unit that performs a posture correction amount in which the position and posture of the affected part are the same a predetermined template region generated from one of the aforementioned three-dimensional reference artifacts or one of the aforementioned three-dimensional current artifacts according to the result of the aforementioned one-time pattern matching, from the predetermined template region with respect to the result according to the aforementioned one-time pattern matching The basis of the generation is different from the aforementioned three-dimensional reference image or the other of the aforementioned three-dimensional image A quadratic pattern matching of the predetermined search target area. 2. The image matching device according to the first aspect of the invention, wherein the verification processing unit includes: generating a three-dimensional region from the three-dimensional reference image based on the affected part information prepared in the three-dimensional reference image The reference template area generation unit of the reference image template area. 3. As claimed in claim 1, the s 1 323646 201249496 has: a human body database input unit for obtaining a human body database from the database device; and an average template region generating unit, The average template region is generated from the organ portion corresponding to the affected part of the patient in the human body database, and the verification processing portion has: performing pattern matching from the aforementioned flat region with respect to the aforementioned three-dimensional reference image And, based on the result of the material pattern matching, a reference template region generating portion of the reference image template region of the three-dimensional region is generated from the three-dimensional reference image. 4. In the case of the image checking device described in claim 2 or 3, the aforementioned check-up portion performs the above-described three-dimensional current image from the aforementioned reference image region at the time of the above-mentioned primary pattern matching. The type matches. 5. The image checking device according to the fourth aspect of the invention, wherein the first checking unit generates a region in which the correlation value with the reference image template region is the highest from the two-dimensional current artifact. The one-time extraction as the search target area appears in the image area. 6. The image collation processing unit according to claim 5, wherein the collation processing unit includes a position and posture converting unit that converts a position and a posture of the three-dimensional imaging image, and the position and posture converting unit generates The position and posture change template region in which the position and orientation of the reference image pattern 2 323646 201249496 is changed to a predetermined position and posture, and the secondary collation portion is subjected to the predetermined template as the predetermined pattern The aforementioned positional position changing template area of the area matches the pattern appearing in the imaging area with respect to the aforementioned one-time drawing. 7. The image checking device according to claim 6, wherein the secondary collating portion generates a cross section of the position and posture changing template region, and then performing the first drawing in the image area and the cross section. Match the pattern between them. 8. The image matching device according to claim 2, wherein the verification processing unit includes a position and orientation conversion unit that changes a position and orientation of the three-dimensional imaging image, and the position and posture conversion unit generates Converting the position and orientation of the reference image template region to a position and posture transformation template region of a predetermined position and posture, and the primary verification portion performs the change from the position and posture to the three-dimensional present state when the primary pattern matching is performed The type of the image matches. 9. The image checking device according to the above claims, wherein the first checking portion generates a cross section of the position and posture changing template region, and then performs between the three-dimensional current image and the aforementioned cross section. Model matching, and execution: the determination of the high correlation section of the section with the highest value of 3 323646 201249496 among the plurality of sections, the calculation of the posture change amount of the position and posture transformation template area, and the aforementioned three-dimensional present portrait Extraction of the extracted area corresponding to the aforementioned high correlation section. 10. The image collation device according to claim 9, wherein the one-time collation unit generates the extracted region extracted by the aforesaid-sub-type matching as the predetermined template. In the current image template area of the area, the position and posture conversion unit generates a three-dimensional posture conversion reference image in which the position and posture of the three-dimensional image are only converted and the current image template area is changed: the posture change amount of the extraction area The secondary collation unit matches the pattern of the three-dimensional posture-converted reference image region as the search target region in the above-described quadratic pattern matching. η. The image material set according to the 1Gth item of the patent scope of the patent, wherein the gold is used by the two persons to check the β system to generate the check surface of the section belonging to the posture transformation reference domain, and the current image of the current image The pattern area matches the pattern between the aforementioned verification execution faces. 12· a patient positioning device having: an eye-catching device as described in any of the 7-11 items; and, according to the foregoing The calculated position correction amount, = the treatment table control to control the light parameters of the treatment table. S 323646 4 201249496 13. The method of checking the image, the three-dimensional image taken during the treatment of the radiation therapy The three-dimensional current image captured by the reference image and the treatment includes: performing a pattern matching step of matching from the aforementioned three-dimensional reference image with respect to the one-dimensional pattern of the aforementioned three-dimensional current image; and performing from the foregoing The predetermined template region generated from one of the aforementioned three-dimensional reference artifacts or one of the aforementioned three-dimensional current artifacts as a result of the pattern matching is different from the generation basis of the predetermined template region with respect to the result of the matching according to the aforementioned one-time pattern A quadratic pattern matching step of the quadratic pattern matching of the predetermined search target region generated by the other of the three-dimensional reference image or the other three-dimensional current image. 14. The method of checking an image according to claim 13, wherein the method includes: generating a reference image template region of the reference image region of the three-dimensional region from the three-dimensional reference image, and the foregoing The one-time pattern matching step performs a pattern matching from the aforementioned reference image template area with respect to the aforementioned three-dimensional current image. 15. A patient positioning device comprising: an image checking device according to claim 4; and calculating a body position correction amount calculated by the image checking device to calculate each of the control stations The treatment table control parameter calculation unit of the parameters of the axis. 16. A patient positioning device comprising: an image checking device according to item 8 of the patent application scope; and 5 323646 201249496 calculating the position correction amount calculated according to the aforementioned image checking device to calculate a treatment table The treatment table control parameter calculation unit of the parameters of each axis. 6 323646