CN1668106A - Method and device for removing block effect of image frame by loop filtering - Google Patents

Abstract

The invention provides an image processing method and a related device for removing the blocking effect of an image picture by loop filtering, which are used for processing the blocking effect between two blocks in the image picture by the loop filtering. The invention compares two boundary pixels at two sides of the boundary of the two blocks according to at least one sensing threshold value to determine whether to adjust the pixel data of the two boundary pixels so as to reduce the difference corresponding to the pixel data of the two boundary pixels.

Description

Method and device for removing block effect of image frame by loop filtering

technical field

The present invention provides a method and a device for processing the block effect of an image frame, especially a method and a device for processing the block effect of an image frame by perceptual threshold loop filtering.

Background technique

Several currently prevailing image coding standards encode an image picture in units of blocks, and the data of each block usually needs to be processed by quantization operations to improve the data compression rate of the encoding process, so the image picture is decoded The block boundary of the block usually has a discontinuity phenomenon, the so-called block effect (blocking artifact). In order to ease the discontinuity caused by image coding and improve image quality, the latest video signal coding standard——H.264 uses a loop filter to deal with the block effect of image frames. For relevant data, please refer to the document of the H.264 video signal coding standard.

The above-mentioned loop filter is a filter set in the encoding loop or decoding loop, which has better processing efficiency than the post-stage filter arranged outside the encoding loop or decoding loop, and does not need to be set as the post-stage filter additional buffer. FIG. 1 is a schematic diagram of a combination of an image processing system 110 , 130 with a loop filter and a transmission/storage medium 120 . The image processing system 110 is an encoding circuit 110 for encoding the image data input from the input terminal 111 . The transmission/storage medium 120 is used to transmit or store the encoded video signal generated by the encoding circuit 110 . The image processing system 130 is a decoding loop 130 for decoding the encoded video signal input from the transmission/storage medium 120 and outputting the decoded image data at the output terminal 133 . The transmission/storage medium 120 may be a transmission channel such as the Internet, or a storage device such as a CD or DVD. The encoding loop 110 has an encoding unit 112 , a reconstruction unit 114 , and a loop filter 116 . The decoding loop 130 has a decoding unit 132 and a loop filter 136 . According to the image processing requirements of the MPEG standard, the predictive frame (P Frame, Predictive Frame) with partial image information must be compared to the intra-coded frame (I Frame, Intra Frame) with complete image information through the encoding loop 110 or decoding loop 130 , to complete the encoding or decoding of the predicted frame. The loop filters 116 and 136 can complete the filtering tasks assigned by the loop filters when the encoding loop 110 or the decoding loop 130 performs encoding or decoding operations, so their processing efficiency is better than that of the post-stage filters. processing efficiency.

Compared with post-stage filtering, although the H.264 video signal coding standard has inherent advantages of loop filtering, the computational complexity has become the bottleneck of the H.264 video signal coding standard. Taking a decoder conforming to the H.264 video signal encoding standard as an example, the loop filter used to remove blocking artifacts accounts for about 33% of the workload of the entire decoding loop.

Contents of the invention

Therefore, the main object of the present invention is to provide a method and device for processing the block effect of image frames by in-loop filtering in the form of perceptual threshold, so as to solve the above-mentioned problems.

The invention provides an image processing method for processing the blocking artifact between two blocks in an image frame. The image processing method includes: storing pixel data corresponding to the two blocks; and comparing two boundary pixels (boundary edge pixels) on both sides of the boundary of the two blocks according to a first threshold to determine whether to adjust the pixels of the two boundary pixels If the difference corresponding to the pixel data of the two border pixels meets the first threshold, adjust the pixel data of the two border pixels to reduce the difference corresponding to the pixel data of the two border pixels.

While providing the above method, the present invention also correspondingly provides a loop filter of an image processing system for processing the block effect between two blocks in an image frame. The loop filter has: a storage unit, used to store pixel data corresponding to the two blocks; a comparison unit, electrically connected to the storage unit, used to compare the two sides of the boundary of the two blocks according to a first threshold Two boundary pixels (boundary edge pixel) to determine whether to adjust the pixel data of the two boundary pixels, if the difference corresponding to the pixel data of the two boundary pixels meets the first threshold, the comparison unit decides to adjust the pixels of the two boundary pixels data to reduce the difference corresponding to the pixel data of the two boundary pixels; and an arithmetic unit (arithmetic unit), electrically connected to the comparison unit and the storage unit, for adjusting the pixel data of the two boundary pixels.

One of the advantages of the present invention is that the present invention uses in-loop filtering to process the block effect of the image frame. Compared with the post-stage filtering, the in-loop filtering has better processing performance.

Another advantage of the present invention is that the present invention decides whether to adjust the pixel data of the two boundary pixels in the form of sensory threshold, so it can quickly decide not to process the block effect that is difficult for human eyes to improve the processing of the image frame efficacy.

Description of drawings

FIG. 1 is a schematic diagram of a combination of a conventional image processing system and a transmission/storage medium.

FIG. 2 is a schematic flow chart of the image processing method of the present invention.

FIG. 3 is a schematic flowchart of intra coding filtering of the method in FIG. 2 .

FIG. 4 is a schematic flow chart of the inter-method (inter) encoding and filtering in FIG. 2 .

FIG. 5 is a schematic diagram of related blocks processed by the method of FIG. 2 .

FIG. 6 is a schematic diagram of the pixel sequence in FIG. 2 .

FIG. 7 is a schematic diagram of pixel data processed by the method of FIG. 2 .

FIG. 8 is a boundary strength comparison table of the method in FIG. 4 .

FIG. 9 is a schematic diagram of the sensory loop filter of the present invention.

FIG. 10 is a schematic diagram of an image coding system applying the perceptual loop filter of FIG. 9 .

FIG. 11 is a schematic diagram of an image decoding system applying the perceptual loop filter of FIG. 9 .

Description of reference symbols 110, 130, 700, 800 image processing system 111, 113, 131, 133 input/output terminals 112 code units 114 Reconstruction Units 116, 136, 600 loop filter 120 transmission/storage media 132 decoding units 300 macro blocks 301-308, 401 Boundary Blocks 315-348 610 storage units 620 comparison unit 630 computing unit m,n vertical vector p _i , q _i pixel data

Detailed ways

Please refer to FIG. 2, FIG. 3, and FIG. 4 at the same time. FIG. 2, FIG. 3, and FIG. 4 are all schematic flow charts of the image processing method of the present invention, wherein FIG. 3 and FIG. 4 show steps 201a and 201b of FIG. 2 respectively. Detailed process. The method of the present invention firstly determines the frame type in step 200, and the step 200 is well known in the art. When a frame currently to be processed is an intra frame (Intra frame), step 201a is performed; otherwise, the frame to be processed currently is an encoded frame (Inter frame), and step 201b is performed. The above-mentioned types of intra-coded frames include: intra-coded slice (Intra slice) and synchronized intra-coded slice (SI slice, Synchronized Intra slice). The above-mentioned types of coded frames include: predictive coded slice (Pslice, Predicted slice), bidirectional predictive coded slice (Bslice, Bidirectional predicted slice), and synchronous predictive coded slice (SP slice, Synchronized Predicted slice). Since the content of step 201b is similar to part of the details of step 201a, step 201a will be described first and then step 201b will be described below.

Please refer to FIG. 3 , FIG. 4 , FIG. 5 , and FIG. 6 at the same time. FIG. 5 is a schematic diagram of related blocks processed by the methods of FIG. 3 and FIG. 4 , and FIG. 6 is a schematic diagram of the pixel sequence in FIG. 3 . In this embodiment, the image frames processed by the methods shown in FIG. 3 and FIG. 4 are composed of macroblocks (macroblock) 300 shown in FIG. 5 , and each macroblock 300 has sixteen blocks (blocks). ) 315, 316, . . . , 348, wherein each block has 4×4 luminance pixel data or 2×2 chromatic pixel data. The vertical axis of Fig. 6 represents the size of the pixel data, and the horizontal axis of Fig. 6 represents a vertical vector n, wherein the vertical vector n is perpendicular to two adjacent blocks P, Q in the image frame (not marked in the relevant figure shown) the boundary 401.

The pixel data p _i , q _i (i=0, 1, . . . ) shown in FIG. 3 , FIG. 4 , and FIG. 6 respectively correspond to the two adjacent blocks P, Q, wherein the pixel data p ₀ and q ₀ respectively represent the data of the boundary pixel (boundary edge pixel) of the two blocks P and Q arranged on the vertical vector n closest to the boundary 401 of the two blocks P and Q, and the pixel data p ₁ , q ₁ Then respectively represent the data of the interior edge pixel (interior edge pixel) arranged on the vertical vector n which is next to the boundary 401 , and so on. For example, the blocks 326 and 336 in FIG. 5 are used as the two adjacent blocks P and Q, and the vertical vector m shown in FIG. 5 is one of the vertical vectors perpendicular to the boundary 303, then the boundary 401 shown in FIG. 6 is The boundary 303 shown in FIG. 5 and the vertical vector n shown in FIG. 6 are the vertical vector m shown in FIG. 5 . At this time, the pixel data p ₀ , p ₁ , . data, and the pixel data q ₀ , q ₁ , . The data. With the selection of different vertical vectors m, the borders 301, 302, . . . , of each block 315, 316, . The blocking artifacts on 308 can be processed one by one by the image processing methods shown in FIG. 3 and FIG. 4 to remove them.

As shown in FIG. 3 , the present invention provides an image processing method for processing the block effect between two blocks P and Q in an image frame. Wherein the image processing method is a video signal encoding process or a loop filtering method in the video signal decoding process. The order of the following steps does not limit the scope of the present invention, and the details of step 201a of the image processing method are described as follows.

Step 202: Store pixel data p _i , q _i (i=0, 1, 2) corresponding to the two blocks P, Q;

Step 204: According to a noticeable difference threshold (Noticeable Difference Threshold) ΔI, compare the pixel data p 0 and _{q 0 of} the two boundary pixels on both sides of the boundary 401 of the two blocks P and Q to determine whether to adjust the pixel data of the two boundary pixels p ₀ , q ₀ , if the difference p ₀ -q ₀ corresponding to the pixel data p ₀ , q ₀ of the two boundary pixels is less than the significant difference threshold ΔI, enter the early termination (Early Termination) state 290 to save time for subsequent processing Otherwise, enter step 206, wherein the significant difference threshold ΔI can also be called critical significant difference (JND, Just Noticeable Difference) ΔI;

Step 206: According to a Recognizable Discontinuity Threshold (Recognizable DiscontinuityThreshold) T(ΔQ, ΔI), compare the pixel data p 0 , q ₀ of the two boundary pixels on both sides of the boundary 401 of the two blocks P, _Q to decide whether to adjust the For the pixel data p ₀ and q ₀ of the two boundary pixels, if the difference p ₀ -q ₀ corresponding to the pixel data p ₀ and q ₀ of the two boundary pixels is less than the distinguishable discontinuity threshold T(ΔQ, ΔI), enter Step 208, otherwise, enter the early termination state 290 to save time to process subsequent pixel data, wherein the recognizable discontinuity threshold T(ΔQ, ΔI) can also be called the recognizable discontinuity limit (Recognizable Discontinuity Limit) T(ΔQ, ΔI );

Step 208: According to an adjustment threshold (Adjustment Threshold) Δ ₀ /2, compare the pixel data p ₀ or q ₀ of a pixel among the two boundary pixels with the pixel data p ₁ or q ₁ of an adjacent internal pixel of the boundary to determine whether Adjust the pixel data p ₀ and q ₀ of the two border pixels, if the pixel data difference p ₁ -p ₀ of the block P is less than the adjustment threshold Δ ₀ /2 or the pixel data difference q ₁ -q ₀ of the block Q is less than Adjust threshold Δ ₀ /2, then enter step 210, otherwise enter early termination state 290 to save time to process subsequent pixel data;

Step 210: Adjust the pixel data p ₀ and q ₀ of the two boundary pixels to reduce the difference p 0 -q ₀ corresponding to the pixel data p ₀ and q ₀ of the two boundary pixels, wherein the difference p _{0 -q 0} is Luminance difference or chromatic difference. The adjustment results of this step are as follows:

p ₀ '=p ₀ +kΔ ₀

q ₀ '=q ₀ -kΔ ₀

And execute steps 212p and 212q respectively after adjustment;

Step 212p: Compare the pixel data p ₀ ' of an adjusted boundary pixel among the two boundary pixels with the pixel data p ₁ of an adjacent boundary internal pixel according to the significant difference threshold ΔI to determine whether to adjust the pixel data of the boundary internal pixel p ₁ , if the difference p ₁ -p ₀ ' is less than the significant difference threshold ΔI, enter the early termination state 291p to save time for processing subsequent pixel data, otherwise enter step 214p;

Step 214p: Calculate the expected adjustment value p ₁ ′=p ₁ +0.5kΔ ₀ of the pixel data of the pixel immediately inside the boundary, and compare the pixel data p ₁ of the pixel inside the boundary with its expected adjustment value p ₁ ’ whichever is closer to the The average value p _m of the adjusted pixel data p ₀ ' of the boundary pixel and the pixel data p ₂ of a pixel immediately adjacent to the boundary is used to determine whether to adjust the pixel data p ₁ of the pixel inside the boundary, if the difference p ₁ '- If p _m is less than p ₁ -p _m , go to step 216p, otherwise the pixel data p ₁ will not be updated. As shown in Figure 7, the average value p _m is the midpoint between pixel data p ₀ ' and pixel data p ₂ ;

Step 216p: Adjust the pixel data p ₁ of the pixels inside the boundary to its expected adjustment value p ₁ '=p ₁ +0.5kΔ ₀ ;

Step 212q: According to the significant difference threshold ΔI, compare the pixel data q ₀ ' of an adjusted boundary pixel among the two boundary pixels with the pixel data q ₁ of an adjacent boundary internal pixel to determine whether to adjust the pixel data of the boundary internal pixel q ₁ , if the difference q ₁ -q ₀ ′ is less than the significant difference threshold ΔI, enter the early termination state 291q to save time for processing subsequent pixel data, otherwise enter step 214q;

Step 214q: Calculate the expected adjustment value q ₁ ′=q ₁ −0.5kΔ ₀ of the pixel data of the pixel immediately inside the boundary, and compare the pixel data q ₁ of the pixel inside the boundary with its expected adjustment value q ₁ ’ whichever is closer to the The average value q _m of the adjusted pixel data q ₀ ' of the boundary pixel and the pixel data q ₂ of a pixel immediately adjacent to the boundary is used to determine whether to adjust the pixel data q ₁ of the pixel inside the boundary. If the difference q ₁ '- q _m is less than q ₁ -q _m , then enter step 216q, otherwise the pixel data q ₁ will not be updated; and

Step 216q: adjust the pixel data q ₁ of the pixels inside the boundary to its expected adjustment value q ₁ '=q ₁ -0.5kΔ ₀ ;

The above significant difference threshold ΔI is defined according to Weber's Law. Weber's law states that the ratio of the increment threshold to the background intensity is constant. In this embodiment, the average brightness of the two blocks P and Q can be defined as I _p and I _q respectively. According to Weber's law, when the background brightness is I _p , the ratio of the critical significant difference ΔI _p that is almost indistinguishable by human eyes to the brightness I _p of the environment is constant at a constant value k. Similarly, when the background brightness is I _q , the ratio of the critical significant difference ΔI _q that is almost indistinguishable by the human eye to the ambient brightness I _q is always a constant value k. As shown in the parameter definition in Figure 3, the significant difference threshold ΔI used in the method of the present invention is the average value (ΔI _p +ΔI _q )/2 of the critical significant difference ΔI _p and ΔI _q of the two blocks P and Q . That is to say, the significant difference threshold ΔI=(ΔI _p +ΔI _q )/2=(kI _p +kI _q )/2. As mentioned above, the method of the present invention is not only applicable to the pixel data p _i , q _i of brightness, but also applicable to the pixel data p _i , q _i of color. Therefore, when the pixel data p _i and q i processed by the method of the present invention are color pixel data p i _{and q i ,} the average values of the pixel data of the two blocks P and Q are respectively defined as I _p and I _q , and then calculate the significant difference threshold ΔI=(ΔI _p +ΔI _q )/2=(kI _p +kI _q )/2 according to the aforementioned definition.

As described in step 216p, the adjustment amount _0.5kΔ0 of the pixel data _p1 of the pixel inside the boundary is half of the adjustment amount _kΔ0 of the pixel data _p0 of the boundary pixel described in step 210. As described in step 216q, the adjustment amount _−0.5kΔ0 of the pixel data _q1 of the pixel inside the boundary is half of the adjustment amount _−kΔ0 of the pixel data _q0 of the boundary pixel described in step 210. In particular, it should be noted that although the parameter definition in Figure 3 shows that the distinguishable discontinuity threshold T(ΔQ, ΔI) is defined as the difference ΔQ and the significant difference threshold ΔI of the quantization parameters (Quantization Parameter) of the two blocks P and Q The linear combination of (αΔQ+βΔI), and in this embodiment can be implemented by using the simplest parameter α=β=1, which does not limit the scope of the present invention. In another embodiment of the present invention, the distinguishable discontinuity threshold T(ΔQ, ΔI) can also be formed by a higher-order polynomial of ΔQ and ΔI or other types of functions of ΔQ and ΔI, wherein when the two When the difference ΔQ or the critical significant difference ΔI of the quantization coefficients of blocks P and Q increases, the distinguishable discontinuity threshold T(ΔQ, ΔI) also increases correspondingly, and when the quantization coefficients of the two blocks P and Q When the difference ΔQ or the critically significant difference ΔI decreases, the discernible discontinuity threshold T(ΔQ, ΔI) also decreases correspondingly. Therefore, the method of the present invention also includes: when the difference ΔQ or the critical significant difference ΔI of the quantization coefficients of the two blocks P and Q increases, increasing the distinguishable discontinuity threshold T(ΔQ, ΔI); When the difference ΔQ or the critically significant difference ΔI of the quantization coefficients of blocks P and Q decreases, the discernible discontinuity threshold T(ΔQ, ΔI) is decreased.

Please refer to Figure 2, Figure 4, and Figure 8 at the same time. Figure 8 is the boundary strength (BS, BoundaryStrength) comparison table of the method in Figure 4. The boundary strength comparison table is well known in the industry, and its definition is similar to that of JVT (H.264 ) standard defined. As shown in FIG. 2 , when the current frame to be processed is an encoded frame, step 201b is performed. However, at the beginning of executing step 201b, first check whether the boundary strength BS of the frame to be processed is zero according to the boundary strength comparison table shown in FIG. 201b, otherwise the pending frame will not be updated. As shown in FIG. 4, the details of step 201b of the image processing method are described as follows.

Step 202: Store pixel data p _i , q _i (i=0, 1, 2) corresponding to the two blocks P, Q;

Step 206': compare the pixel data p 0 , q 0 of the two boundary pixels on both sides of the boundary 401 of the two blocks P, Q according to a threshold T to determine whether to adjust the pixel data _{p 0 , q 0 of} the two boundary pixels, If the difference p ₀ -q ₀ corresponding to the pixel data p ₀ and q ₀ of the two boundary pixels is less than the threshold T, then enter step 208 ′, otherwise enter the early termination state 290 to save time for processing subsequent pixel data. Wherein the definition of the threshold T of this step is the parameter definition as shown in Figure 4;

Step 208': Compare the pixel data _p0 or q0 of a pixel of the two boundary pixels with the pixel data _p1 or _q1 of an adjacent internal pixel according to an _adjustment threshold _Δ0 /2 to determine whether to adjust the two For the pixel data p ₀ and q ₀ of the boundary pixels, if the pixel data difference p ₁ -p ₀ of the block P is less than the adjustment threshold Δ ₀ /2 or the pixel data difference q ₁ -q ₀ of the block Q is less than the adjustment threshold Δ _0/2 , then enter step 210', otherwise enter early termination state 290 to save time to process subsequent pixel data. Wherein the definition of parameter Δ ₀ is as shown in Figure 4, so the definition of the adjustment threshold Δ _0/2 of this step also changes correspondingly;

Step 210': Adjust the pixel data p ₀ and q ₀ of the two boundary pixels to reduce the difference p 0 -q ₀ corresponding to the pixel data p ₀ and q ₀ of the two boundary pixels, wherein the difference p _{0 -q 0} For brightness difference or color difference. The adjustment results of this step are as follows:

p ₀ '=p ₀ +kΔ ₀

q ₀ '=q ₀ -kΔ ₀

The definition of the parameter _Δ0 is shown in FIG. 4, so the adjustment result of this step is changed accordingly.

In this embodiment, when the average value of the quantization coefficients of the two blocks P and Q is less than 16, the block effect of the boundary 401 of the two blocks P and Q will not be processed to save time. Blocking effects on other borders of the image frame.

Please refer to FIG. 9, FIG. 10, and FIG. 11 at the same time. FIG. 9 is a schematic diagram of the sensory loop filter 600 of the present invention, and FIG. 10 and FIG. 11 are respectively a schematic diagram of an image processing system using the sensory loop filter 600 of FIG. 9 . While providing the above method, the present invention also correspondingly provides a loop filter 600 of an image processing system for processing the block effect between two blocks P and Q in an image frame. The image processing system is a video signal encoder 700 or a video signal decoder 800 . As described above, since the present invention uses the significant difference threshold ΔI and the distinguishable discontinuity threshold T(ΔQ, ΔI) related to whether human eyes can distinguish, the loop filter 600 can also be called a sensory loop filter ( PLF, Perceptual Loop Filter) 600. The loop filter 600 includes: a storage unit 610 for storing pixel data p _i and q _i corresponding to the two blocks P and Q; The difference threshold ΔI compares the pixel data p 0 , _{q 0} of the two boundary pixels on both sides of the boundary 401 of the two blocks P, Q to determine whether to adjust the pixel data p ₀ , _{q 0 of} the two boundary pixels. If the difference p ₀ -q ₀ corresponding to the pixel data p ₀ and q ₀ of the two boundary pixels is less than the significant difference threshold ΔI, the comparison unit 620 decides to enter the early termination state 290 as described in step 204 to save time for processing Subsequent pixel data, otherwise, as described in step 206, the comparison unit 620 compares the pixel data _p0 , q ₀ to determine whether to adjust the pixel data p ₀ , q ₀ of the two boundary pixels. If the difference p ₀ -q ₀ corresponding to the pixel data p ₀ and q ₀ of the two boundary pixels is smaller than the distinguishable discontinuity threshold T(ΔQ, ΔI), the comparison unit 620 decides to perform further comparison as in step 208, otherwise The comparison unit 620 decides to enter the early termination state 290 to save time for processing subsequent pixel data. As mentioned above, the pixel data p ₀ , q ₀ may be pixel data p ₀ , q _{0 of brightness, or pixel data p 0 , q 0 of} color. Therefore, the above-mentioned difference p ₀ -q ₀ may be a brightness difference or a color difference.

The loop filter 600 further includes: an arithmetic unit 630 electrically connected to the comparison unit 620 and the storage unit 610 for adjusting the pixel data p _i , q _i of the two boundary pixels. As described in step 208, the comparison unit 620 also compares _the pixel data _p0 or q0 of a pixel of the two boundary pixels with the pixel data _p1 or _q1 of an adjacent internal pixel of the boundary according to an adjustment threshold _Δ0 /2 to determine whether to adjust the pixel data p ₀ and q ₀ of the two boundary pixels. If the pixel data difference p ₁ -p ₀ of the block P is less than the adjustment threshold Δ ₀ /2 or the pixel data difference q ₁ -q ₀ of the block Q is less than the adjustment threshold Δ ₀ /2, then the comparison unit 620 decides as follows: As described in 210, the operation unit 630 adjusts the pixel data p ₀ and q ₀ of the two boundary pixels to reduce the difference p ₀ -q ₀ corresponding to the pixel data p ₀ and q ₀ of the two boundary pixels. The comparison unit 620 can also perform steps 212zp and 214p to determine whether to adjust the pixel data p ₁ of the pixels inside the boundary to its expected adjustment value p ₁ '=p ₁ +0.5kΔ ₀ by the operation unit 630 as described in step 216p . As mentioned above, the adjustment amount _0.5kΔ0 of the pixel data _p1 of the pixel inside the boundary is half of the adjustment amount _kΔ0 of the pixel data _p0 of the boundary pixel described in step 210. Similarly, the comparison unit 620 can also execute steps 212q and 214q to determine whether to adjust the pixel data q ₁ of the pixels inside the boundary to its expected adjustment value q ₁ ′=q ₁ − 0.5kΔ ₀ . As mentioned above, the adjustment amount _−0.5kΔ0 of the pixel data _q1 of the pixels inside the boundary is half of the adjustment amount _−kΔ0 of the pixel data _q0 of the boundary pixels described in step 210.

In this embodiment, when the difference ΔQ or the significant difference threshold ΔI between the quantization coefficients of the two blocks P and Q increases, the comparison unit 620 increases the distinguishable discontinuity threshold T(ΔQ, ΔI); when the two blocks When the difference ΔQ or the significant difference threshold ΔI between the quantization coefficients of P and Q decreases, the comparing unit 620 decreases the distinguishable discontinuity threshold T(ΔQ, ΔI).

As described above, the difference between the elements in this embodiment when executing 206', 208', and 210' is only in the selection of the threshold T and the parameter _Δ0 , which will not be repeated here.

In the quantization operation process of general image processing, the larger the quantization coefficient selected for each block, the worse the image quality; the smaller the quantization coefficient, the better the image quality. The range of quantization coefficients corresponding to better image quality is about 22 or less. The method and device of the present invention can achieve the same quality of the processed image as that processed by the loop filter conforming to the H.264 video signal coding standard under the condition that the quantization coefficient is less than 22.

One of the advantages of the present invention is that the present invention uses in-loop filtering to process the block effect of the image frame. Compared with the post-stage filtering, the in-loop filtering has better processing performance.

Another advantage of the present invention is that the present invention decides whether to adjust the pixel data of the two boundary pixels in the form of sensory threshold, so it can quickly decide not to process the block effect that is difficult for human eyes to improve the processing of the image frame efficacy.

Another benefit of the present invention is that the calculation of the method and related devices of the present invention is simple and simple, and in terms of the image frame formed by the blocks arranged in the horizontal and vertical directions, one side of the border of each block is connected to the area. Only two pixels of pixel data in the vertical direction of the block boundary need to be adjusted, so the processing performance is better than that of the conventional H.264 standard.

The above descriptions are only preferred embodiments of the present invention, and all equivalent changes and modifications made according to the claims of the present invention shall fall within the scope of the present invention.

Claims (20)

1. an image processing method is used for handling the block effect between two blocks in the image frame, and this image processing method includes:

Storage is to pixel data that should two blocks; And

According to a first threshold relatively two border pixels of the boundaries on either side of this two block determine whether adjusting the pixel data of this two borders pixel, if the pairing difference of pixel data of this two borders pixel meets this first threshold, the pixel data of adjusting this two borders pixel is to reduce the pairing difference of pixel data of this two borders pixel.

2. the method for claim 1, wherein this difference is a luminance difference XOR heterochromia.

3. the method for claim 1, it also includes: the pixel data that determines whether adjusting this two borders pixel according to one second threshold ratio than a pixel in this two borders pixel and a next-door neighbour's border interior pixels.

4. method as claimed in claim 3, it also includes: the pixel data that determines whether adjusting this border interior pixels according to one the 3rd threshold ratio than an adjusted boundary pixel in this two borders pixel and a next-door neighbour's border interior pixels.

5. method as claimed in claim 4, wherein the 3rd threshold value equals this first threshold.

6. method as claimed in claim 4, wherein the pixel data adjustment amount of this border interior pixels is half of pixel data adjustment amount of this boundary pixel.

7. the method for claim 1, it also includes:

When the difference of the quantization parameter of this two block increases, increase this first threshold; And

When the difference of the quantization parameter of this two block reduces, reduce this first threshold.

8. the method for claim 1, it also includes:

When one increases according to the defined critical significant difference of Weber's law, increase this first threshold; And

When this critical significant difference reduces, reduce this first threshold.

9. the method for claim 1, wherein this first threshold is one according to the defined critical significant difference of Weber's law.

10. the method for claim 1, wherein this image processing method is the loop filter method of a video signal coding process or vision signal decode procedure.

11. the loop filter of an image processing system is used for handling the block effect between two blocks in the image frame, this loop filter includes:

One memory cell is used for storing to pixel data that should two blocks;

One comparing unit, be electrically connected to this memory cell, be used for according to a first threshold relatively two border pixels of the boundaries on either side of this two block determine whether adjusting the pixel data of this two borders pixel, if the pairing difference of pixel data of this two borders pixel meets this first threshold, this comparing unit decision is adjusted the pixel data of this two borders pixel to reduce the pairing difference of pixel data of this two borders pixel; And

One arithmetic element is electrically connected to this comparing unit and this memory cell, is used for adjusting the pixel data of this two borders pixel.

12. loop filter as claimed in claim 11, wherein this difference is a luminance difference XOR heterochromia.

13. loop filter as claimed in claim 11, wherein this comparing unit also determines whether adjusting the pixel data of this two borders pixel than a pixel in this two borders pixel and a next-door neighbour's border interior pixels according to one second threshold ratio.

14. loop filter as claimed in claim 13, wherein this comparing unit also determines whether adjusting the pixel data of this border interior pixels than an adjusted boundary pixel in this two borders pixel and a next-door neighbour's border interior pixels according to one the 3rd threshold ratio.

15. method as claimed in claim 14, wherein the 3rd threshold value equals this first threshold.

16. loop filter as claimed in claim 14, wherein the pixel data adjustment amount of this border interior pixels is half of pixel data adjustment amount of this boundary pixel.

17. loop filter as claimed in claim 11, wherein when the difference of the quantization parameter of this two block increased, this comparing unit increased this first threshold; When the difference of the quantization parameter of this two block reduced, this comparing unit reduced this first threshold.

18. loop filter as claimed in claim 11, wherein when one increased according to the defined critical remarkable difference of Weber's law, this comparing unit increased this first threshold; When this critical significant difference reduced, this comparing unit reduced this first threshold.

19. loop filter as claimed in claim 11, wherein this first threshold is one according to the defined critical significant difference of Weber's law.

20. loop filter as claimed in claim 11, wherein this image processing system is a video coder or a video signal decoder.

Images