WO2017032190A1 - Image identification method and apparatus - Google Patents

Abstract

An image identification method and apparatus. The method comprises: acquiring an image to be analyzed (S101, S201); identifying a geometrical shape comprised in the image (S102, S203); comparing the geometrical shape with a set reference shape, and determining a target geometrical shape matching the reference shape (S103); and marking and displaying the target geometrical shape (S104). The solution is capable of identifying different geometrical shapes in an image according to different reference shapes, more flexibly abstracting and simplifying an object in the image, and assisting students to intuitively cognize geometric figures in natural life.

Description

Image recognition method and device Technical field

The present invention relates to the field of image recognition technologies, and in particular, to a method and apparatus for shape recognition in an image.

Background technique

In geometry, points, lines and other elements are often used to describe objects. However, when people look at objects with the naked eye, they cannot intuitively feel which geometric shapes are included in the objects.

In the field of teaching, existing equipment or analysis software can identify the lines contained in the image by analyzing the pixels contained in the image. However, this recognition usually only goes to the level of the line, and does not consider the geometric shapes that may be included in the image, and it is impossible to distinguish the different geometric shapes formed by the lines separately.

Therefore, the prior art in the teaching field is not ideal for abstracting and simplifying real objects, and it is impossible for students to have an intuitive feeling of geometric figures in natural life.

Summary of the invention

Based on this, the present invention provides a method and device for image recognition, which can identify different kinds of geometric figures contained in objects in an image, and abstract and simplify the real objects to a certain extent.

The invention adopts the following technical solutions:

An aspect of the present invention provides a method for image recognition, including:

Obtain an image to be analyzed;

Identifying geometric shapes included in the image;

Aligning the geometric shape with the set reference shape to determine a target geometry that matches the reference shape;

The target geometry is marked for display.

As a preferred manner, the identifying the geometric shape included in the image includes:

Obtaining a gray value of each pixel in the image;

Determining adjacent pixel points whose gray value differs by less than or equal to the set tolerance, and identifying a line formed by the adjacent pixel points;

A closed shape formed by the lines is identified, the closed shape being determined as a geometric shape included in the image.

As a preferred manner, comparing the geometric shape with the set reference shape, determining a target geometry that matches the reference feature comprises:

Extracting a reference feature of the reference shape, and extracting feature information of the geometric shape;

Calculating the feature information, the approximation degree of the reference feature, and if the approximation is greater than or equal to the set threshold, determining that the corresponding geometric shape is the target geometry.

As a preferred manner, after determining the corresponding geometry as the target geometry, the method further includes:

Correcting the target geometry according to the reference feature, so that the corrected target geometry has all the reference features corresponding to the reference shape.

In a preferred manner, the comparing the geometric shape with the set reference shape, determining the target geometry matching the reference feature comprises: setting a reference shape;

The reference shape includes: a circle, a rectangle, a triangle, or a star;

The reference features include: distance information from each point to the center point, number of vertices, angle of vertex information, and/or side length information.

Another aspect of the present invention provides an apparatus for image recognition, including:

An image acquisition module, configured to acquire an image to be analyzed;

a first identification module, configured to identify a geometric shape included in the image;

a second identification module, configured to compare the geometric shape with the set reference shape, and determine a target geometry that matches the reference shape;

A display module is configured to display the target geometry.

As a preferred mode, the first identification module includes:

a pixel point analyzing unit, configured to acquire a gray value of each pixel in the image;

a line identifying unit, configured to determine adjacent pixels whose gray values differ by less than or equal to a set tolerance, and identify lines formed by the adjacent pixels;

And a shape recognition unit for identifying a closed shape formed by the line, the closed shape being determined as a geometric shape included in the image.

In a preferred manner, the second identification module includes:

a feature extraction unit, configured to extract a reference feature of the reference shape, and extract feature information of the geometric shape;

And the feature comparison unit calculates the approximation degree of the feature information and the reference feature, and if the approximation is greater than or equal to the set threshold, determining that the corresponding geometry is the target geometry.

In a preferred embodiment, the method further includes: a correction module, configured to modify the target geometry according to the reference feature, so that the corrected target geometry has all reference features corresponding to the reference shape.

As a preferred manner, the method further includes: a setting module, configured to set a reference shape;

The reference shape includes: a circle, a rectangle, a triangle, or a star;

The reference features include: distance information from each point to the center point, number of vertices, angle of vertex information, and/or side length information.

The beneficial effects of implementing the above technical solution of the present invention include: identifying, based on the set reference shape, whether there is a target geometric shape matching the reference shape in the image to be analyzed, and marking the existing target geometric shape in the image; According to different reference shapes, different types of geometric shapes in the image can be identified, and the objects in the image can be more abstracted and simplified, and the students (users) can intuitively recognize the geometric figures in natural life.

DRAWINGS

FIG. 1 is a schematic flowchart of a method for image recognition according to an embodiment of the present invention; FIG.

2 is a schematic diagram of an image recognition effect;

3 is a schematic diagram of another image recognition effect;

FIG. 4 is a schematic flowchart of a method for image recognition according to another embodiment of the present invention; FIG.

FIG. 5 is a schematic diagram of another image recognition effect;

Figure 6 is a schematic diagram of an image correction effect;

FIG. 7 is a schematic structural diagram of an apparatus for image recognition according to an embodiment of the present invention; FIG.

FIG. 8 is a schematic structural diagram of an apparatus for image recognition according to another embodiment of the present invention.

detailed description

The technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. The embodiments are merely a part of the embodiments of the invention, and not all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

The invention provides a method for image recognition, which is especially suitable for abstracting and simplifying processing of objects in an image in the teaching field, identifying different types of geometric shapes in the image, and assisting students (users) to intuitively construct geometric figures in natural life. Cognition. Embodiments of the present invention also provide corresponding apparatus for image recognition. The details are described below separately.

FIG. 1 is a schematic flowchart of a method for image recognition according to an embodiment of the present invention. As shown in FIG. 1, the method includes the following steps S101 to S104, which are described in detail as follows:

Step S101, acquiring an image to be analyzed;

The image to be analyzed may be a locally stored image, an imported external image, or an image taken instantaneously by a camera. Preferably, only one image is acquired at a time as an image to be analyzed.

Step S102, identifying a geometric shape included in the image;

In this embodiment, the manner of identifying the geometric shape included in the image may be: first processing the image into a gray image, and acquiring gray values of each pixel in the image; determining that the gray value is less than or equal to the set capacity. Poor adjacent pixels, then identifying lines formed by the adjacent pixels; further identifying a closed shape formed by the lines, the closed shape being determined as the geometric shape contained in the image.

It should be noted that the tolerance setting has a key influence on the recognition accuracy and the recognition strength. Generally, the smaller the tolerance setting, the higher the recognition accuracy, but the smaller the recognition strength (that is, the possible lines cannot be recognized. Come out); the larger the tolerance setting, the lower the recognition accuracy, but the higher the recognition strength. Therefore, it is necessary to adjust the tolerance according to the actual situation.

Step S103, comparing the geometric shape with the set reference shape, determining a target geometry matching the reference shape;

Preferably, the specific manner of step S103 may be: firstly, extracting reference features of the set reference shape, and separately extracting feature information of the geometric shape identified by the above steps; and then separately calculating feature information of each geometric shape and the reference The approximation of the feature. The approximation can be expressed as a percentage. The higher the degree of approximation, the more similar the corresponding geometric shape is to the reference shape, and the more features satisfying the reference shape, when the degree of approximation is 100%, the corresponding geometric shape has all the references corresponding to the reference shape. The feature; conversely, the lower the degree of approximation, the greater the difference between the corresponding geometric shape and the reference shape, and the fewer features corresponding to the reference shape.

In this embodiment, according to the set matching determination threshold, if the approximation is greater than or equal to the threshold, it is determined that the corresponding geometry is the target geometry.

It should be noted that the reference shape, such as a circle, a rectangle, a triangle, or a star, needs to be preset before the identification, and the above reference features include, but are not limited to, distance information from each point to the center point, the number of vertices, the angle of the vertex, and / or side length information. For example, a circular reference feature includes lines that are smooth and continuous, and each point-to-center distance is the same; a reference feature of a triangle includes a closed figure composed of three sides and three vertices; a rectangular reference feature includes four sides, four vertices, And all the apex angles are right angles and so on.

Step S104, displaying the target geometry.

Preferably, a specific color may be filled in a corresponding area of the target geometric shape in the image, or a new layer may be created on the image, and a corresponding geometric line drawing may be drawn on the newly created layer, that is, in the layer. The position of the geometric line map is aligned with the position of the corresponding target geometry in the image, and then the new layer is derived to clearly show the target geometry contained in the image.

The above embodiments will be described below with two specific examples.

As shown in FIG. 2, the image to be analyzed is a photo of a bicycle, and the reference shape is set to a triangle, and the feature information of the triangle is extracted, for example, three vertices, and the three vertices are not on the same straight line, and the sum of the vertices is 180 degrees, the sum of the lengths of any two sides is greater than the third side, and the like. It then begins to identify the lines in the bicycle image, resulting in a closed shape of lines, determining whether the features satisfy the characteristics of the triangle, and if so, marking the triangle at the corresponding position in the bicycle image. As shown in Figure 2, the tripod of the bicycle body and the triangle formed by the spokes in the wheel are identified (all triangles are not shown in the figure, just an example), or the identified triangles may be derived, and the identified multiples will be identified. The triangles are displayed separately.

As shown in FIG. 3, the image to be analyzed is a photograph of a bicycle, and the reference shape is reset to a circular shape, and the circular features can be extracted: a continuous smooth closed curve, and the distance from each point to the center point is equal. Start recognition, it can be recognized that the two wheels of the bicycle are round, and then the indicator display, or two rounds Export is displayed separately.

Through the above described identification process, students can clearly see the different geometries in the bicycle and improve the perception of geometric patterns in natural life.

FIG. 4 is a schematic flowchart of a method for image recognition according to another embodiment of the present invention; as shown in FIG. 4, the method includes:

Step S201, acquiring an image to be analyzed;

Step S202, setting a reference shape;

Step S203, identifying a geometric shape included in the image;

Step S204, determining whether the approximate degree of the geometric shape and the reference shape meets the set condition. If yes, go to the next step, otherwise, go back to the previous step.

In this embodiment, the setting condition is a threshold of the set approximation, and the threshold is less than 100%, for example, 80%;

It should be noted that the specific implementation manners of the foregoing steps S201 to S204 can be referred to the foregoing embodiments, and details are not described herein.

Step S205, displaying the identified target geometry in the image.

For example, a transparent layer is newly created on the image, and a corresponding geometric line drawing is drawn on the transparent layer, that is, the position of the geometric line drawing in the transparent layer and the corresponding target geometric shape in the image. The positions are aligned to clearly show the target geometry contained in the image.

Step S206, does the identified target geometry have all the reference features corresponding to the reference shape? If yes, go to step S208, if no, go to the next step.

Step S207, correcting the target geometry according to the reference feature, so that the corrected target geometry has all the reference features corresponding to the reference shape;

It should be noted that, in this embodiment, when the feature information of the comparison geometry and the reference feature of the reference shape determine the target geometry, a certain fault tolerance range may be set, that is, the threshold is set to be less than 100%, and the approximate degree is greater than or equal to The geometry of the threshold can be confirmed as the target geometry. For example, the ellipse can also be recognized as a circle, the trapezoid is also recognized as the target geometry of the rectangular reference shape, or the diamond is also recognized as the target geometry of the square reference shape; in addition, if there are some unexpected protrusions, turns in the line If it is not smooth, it can also be recognized as a smooth line.

In step S208, the corrected geometry is derived and displayed.

The following further illustrates this embodiment with two specific examples.

As shown in FIG. 5, the image to be analyzed is an image of a fence, and the reference shape is set to a rectangle. The two quadrilaterals shown in the figure can be identified through the above steps 201-204. However, the two quadrilaterals are not standard rectangles, and after the quadrilateral is subjected to a certain correction/deformation by the step S207, the four apex angles of the quadrilateral are at right angles, and the standard rectangle can be changed. At the same time, the correction process can be dynamically displayed to further help students understand the deformation between various geometric shapes.

It should be noted that the above-mentioned correction/deformation rule is not arbitrary, but based on the reference feature information of the reference shape, and follows the regular change of the depth of field transformation in the three-dimensional figure.

As shown in Fig. 6, the circular shape is set as the reference shape, and the graphic on the left side is a geometric shape close to a circle, because although one of the points is recessed inward, most of the points satisfy the characteristics of the circle, and a certain fault tolerance. This graphic can be recognized as a circle within the range. Further, the standard circular shape information may be modified according to the standard circular information, including removing the inwardly recessed points in the left geometry, connecting the notched portions so as to satisfy all the features of the standard circle, and correcting the right The side is regular in a circle.

It should be noted that, for the foregoing method embodiments, for the sake of brevity, they are all described as a series of action combinations, but those skilled in the art should understand that the present invention is not limited by the described action sequence, because In accordance with the present invention, certain steps may be performed in other sequences or concurrently. In the following, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

According to the above-described embodiment of the present invention, based on the set reference shape (geometry), it is recognized whether there is a target geometry matching the reference feature of the reference shape in the input image, and the target geometry to be present in the image Marked out; according to different reference shapes, different geometric shapes in the image can be identified, and the objects in the image can be more abstracted and simplified, which is convenient for the students (users) to intuitively recognize the geometric figures in natural life. In addition, according to the characteristics of the graphics, it can be automatically corrected to a more standard graphics, which is conducive to students to distinguish and contrast different geometric shapes.

An apparatus for image recognition of an embodiment of the present invention which can be used to perform the above method of image recognition will be described below. FIG. 7 is a schematic structural diagram of an apparatus for image recognition according to an embodiment of the present invention. For convenience of description, only parts related to the embodiment of the present invention are shown in the figure, and those skilled in the art can understand the structure of the apparatus shown in the figure. Does not constitute a limitation on the device, and may include more or less parts than the illustration. Pieces, or combine some parts, or different parts.

FIG. 7 is a schematic structural diagram of an apparatus for image recognition according to an embodiment of the present invention. As shown in Figure 7, the device comprises:

An image obtaining module 710, configured to acquire an image to be analyzed;

a first identification module 720, configured to identify a geometric shape included in the image;

a second identification module 730, configured to compare the geometric shape with the set reference shape, and determine a target geometry that matches the reference shape;

The display module 740 is configured to display the target geometry.

Preferably, the apparatus for image recognition of the embodiment further includes a setting module 750, configured to set a reference shape;

The above reference shapes include, but are not limited to, a circle, a rectangle, a triangle, or a star; the above reference features include, but are not limited to, distance information from each point to a center point, number of vertices, angle of vertex information, and/or side length information.

As a preferred embodiment, the first identification module 720 illustrated in FIG. 7 may specifically include:

a pixel point analyzing unit 721, configured to acquire a gray value of each pixel in the image;

a line identifying unit 722, configured to determine adjacent pixel points whose gray value differences are less than or equal to a set tolerance, and identify a line formed by the adjacent pixel points;

And, the shape recognition unit 723 is configured to identify a closed shape formed by the line, and determine the closed shape as a geometric shape included in the image.

As a preferred embodiment, the second identification module 730 illustrated in FIG. 7 may specifically include:

a feature extraction unit 731, configured to extract a reference feature of the reference shape, and extract feature information of the geometric shape;

And the feature comparison unit 732 calculates the approximation degree of the feature information and the reference feature, and if the approximation is greater than or equal to the set threshold, determines that the corresponding geometry is the target geometry.

FIG. 8 is a schematic structural diagram of an apparatus for image recognition according to another embodiment of the present invention. As shown in FIG. 8, the apparatus for image recognition of the embodiment further includes a correction module 760, configured to correct the target geometry according to the reference feature, so that the corrected target geometry has the reference shape corresponding to All reference features.

It should be noted that the information interaction between the modules/units in the foregoing device embodiment, the execution process, and the like are based on the same concept as the foregoing method embodiments of the present invention, and the technical effects thereof and the foregoing method embodiments of the present invention are provided. For the details, refer to the description in the method embodiment of the present invention, and details are not described herein again.

In addition, in the implementation of the apparatus for image recognition of any of the foregoing examples of FIG. 7 and FIG. 8 , the logical division of each functional module is merely an example, and may be used in an actual application, for example, according to configuration requirements of the corresponding hardware or software. For the convenience of implementation, the above-mentioned function allocation is completed by different functional modules, that is, the internal structure of the device for image recognition is divided into different functional modules to complete all or part of the functions described above.

According to the image recognition apparatus of the above embodiment, the user can set a reference shape (geometry), and the apparatus based on the image recognition can automatically recognize whether there is a target geometry matching the reference shape in the input image, and The existing target geometry is marked in the image; according to different reference shapes, different geometric shapes in the image can be identified, and the objects in the image are more abstracted and simplified, which is beneficial to the students (users) to nature. The geometry in life has an intuitive perception.

In the above embodiments, the descriptions of the various embodiments are all focused, and the parts that are not detailed in a certain embodiment can be referred to the related descriptions of other embodiments.

In addition, one of ordinary skill in the art will appreciate that all or a portion of the steps of the method specified by any embodiment of the present invention can be accomplished by the program instructing the associated hardware (personal computer, server, or network device, etc.). The program can be stored in a computer readable storage medium. The program, when executed, may perform all or part of the steps of the method specified in any of the above embodiments. The foregoing storage medium may include any medium that can store program code, such as a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

The above is a description of the method and apparatus for image recognition provided by the present invention. For those skilled in the art, according to the idea of the embodiment of the present invention, there will be changes in the specific implementation manner and application scope, in summary, The contents of this specification are not to be construed as limiting the invention.

Claims (10)

A method for image recognition, comprising: Obtain an image to be analyzed; Identifying geometric shapes included in the image; Aligning the geometric shape with the set reference shape to determine a target geometry that matches the reference shape; The target geometry is marked for display. The method of image recognition according to claim 1, wherein said identifying a geometric shape included in said image comprises: Obtaining a gray value of each pixel in the image; Determining adjacent pixel points whose gray value differs by less than or equal to the set tolerance, and identifying a line formed by the adjacent pixel points; A closed shape formed by the lines is identified, the closed shape being determined as a geometric shape included in the image. The method of image recognition according to claim 1, wherein said comparing said geometric shape with said set reference shape to determine a target geometry that matches said reference feature comprises: Extracting a reference feature of the reference shape, and extracting feature information of the geometric shape; Calculating the feature information, the approximation degree of the reference feature, and if the approximation is greater than or equal to the set threshold, determining that the corresponding geometric shape is the target geometry. The method of image recognition according to claim 3, wherein after the determining the corresponding geometric shape as the target geometric shape, the method further comprises: Correcting the target geometry according to the reference feature, so that the corrected target geometry has all the reference features corresponding to the reference shape. The method of image recognition according to claim 3, wherein said comparing said geometric shape with said set reference shape to determine a target geometry that matches said reference feature comprises: Set the reference shape; The reference shape includes: a circle, a rectangle, a triangle, or a star; The reference feature includes: distance information from each point to the center point, number of vertices, and angle of the vertex And / or side length information. An apparatus for image recognition, comprising: An image acquisition module, configured to acquire an image to be analyzed; a first identification module, configured to identify a geometric shape included in the image; a second identification module, configured to compare the geometric shape with the set reference shape, and determine a target geometry that matches the reference shape; A display module is configured to display the target geometry. The apparatus for image recognition according to claim 6, wherein the first identification module comprises: a pixel point analyzing unit, configured to acquire a gray value of each pixel in the image; a line identifying unit, configured to determine adjacent pixels whose gray values differ by less than or equal to a set tolerance, and identify lines formed by the adjacent pixels; And a shape recognition unit for identifying a closed shape formed by the line, the closed shape being determined as a geometric shape included in the image. The apparatus for image recognition according to claim 6, wherein the second identification module comprises: a feature extraction unit, configured to extract a reference feature of the reference shape, and extract feature information of the geometric shape; And the feature comparison unit calculates the approximation degree of the feature information and the reference feature, and if the approximation is greater than or equal to the set threshold, determining that the corresponding geometry is the target geometry. The apparatus for image recognition according to claim 8, further comprising: And a correction module, configured to modify the target geometry according to the reference feature, so that the corrected target geometry has all reference features corresponding to the reference shape. The apparatus for image recognition according to claim 8, further comprising: Setting a module for setting a reference shape; Wherein the reference shape comprises: a circle, a rectangle, a triangle or a star; The reference features include: distance information from each point to the center point, number of vertices, angle of vertex information, and/or side length information.

Images