CN110072034B - Image processing method and image processing device - Google Patents

Image processing method and image processing device Download PDF

Info

Publication number
CN110072034B
CN110072034B CN201810062839.6A CN201810062839A CN110072034B CN 110072034 B CN110072034 B CN 110072034B CN 201810062839 A CN201810062839 A CN 201810062839A CN 110072034 B CN110072034 B CN 110072034B
Authority
CN
China
Prior art keywords
pixel
weight
value
target pixel
reference pixels
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201810062839.6A
Other languages
Chinese (zh)
Other versions
CN110072034A (en
Inventor
刘楷
黄文聪
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Realtek Semiconductor Corp
Original Assignee
Realtek Semiconductor Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Realtek Semiconductor Corp filed Critical Realtek Semiconductor Corp
Priority to CN201810062839.6A priority Critical patent/CN110072034B/en
Publication of CN110072034A publication Critical patent/CN110072034A/en
Application granted granted Critical
Publication of CN110072034B publication Critical patent/CN110072034B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N5/00—Details of television systems
    • H04N5/14—Picture signal circuitry for video frequency region
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N5/00—Details of television systems
    • H04N5/14—Picture signal circuitry for video frequency region
    • H04N5/21—Circuitry for suppressing or minimising disturbance, e.g. moiré or halo

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Image Processing (AREA)

Abstract

The invention discloses an image processing method and an image processing device, which are used for filtering an image. The image processing method comprises the following steps: (A) generating a weight for each reference pixel, wherein the image comprises a plurality of reference pixels, and the size of the weight is related to the similarity between a target pixel and the reference pixel corresponding to the weight; and (B) performing a wireless impulse response filtering operation according to the weights, the pixel value of the target pixel and the pixel values of the reference pixels to obtain a wireless impulse response filtering value of the target pixel.

Description

Image processing method and image processing device
Technical Field
The present invention relates to image processing, and more particularly, to an Infinite Impulse Response (IIR) based image processing method and an image processing apparatus.
Background
FIG. 1 is a diagram of a window for image processing. The 5 × 5 window (containing 25 pixels) in the figure is used to perform image processing on the central target pixel P (i, j), and the other pixels in the window are reference pixels. Assuming that an Infinite Impulse Response (IIR) filtering operation is performed in the spatial domain, the image is processed in a scanning order from top left to bottom right (Raster scan), so that the pixels of the white portion in the window are the original pixels without IIR filtering, and the pixel value thereof is the original pixel value PV ORIThe pixels of the gray portion are pixels that have been previously subjected to the IIR filtering process (referred to simply as IIR pixels), and the pixel values thereof are IIR filtered values PVIIR. In detail, the reference pixel P (i-2, j-2) is an IIR pixel (pixel value is PV)IIR(i-2, j-2)), the reference pixel P (i-2, j-1) is an IIR pixel (pixel value is PV)IIR(i-2, j-1)), …; the target pixel P (i, j) is the original pixel (pixel value is PV)ORI(i, j)), the reference pixel P (i, j +1) is the original pixel (pixel value is PV)ORI(i, j +1)), the pixel P (i, j +2) is the original pixel (pixel value is PV)ORI(i,j+2))、…。
The IIR filtering is to average the IIR pixel in the window with the target pixel to filter the target pixel, and as an example shown in fig. 1, the IIR filtering is to obtain an IIR filtering value PV of the target pixel P (i, j) according to equation (1)IIR(i, j), i.e. calculating the pixel values PV of the 12 IIR pixelsIIRAnd the pixel value PV of the target pixel P (i, j)ORIArithmetic mean of (i, j):
PVIIR(i,j)＝(PVIIR(i-2,j-2)+PVIIR(i-2,j-1)+PVIIR(i-2,j)+PVIIR(i-2,j+1)+PVIIR(i-2,j+2)+PVIIR(i-1,j-2)+PVIIR(i-1,j-1)+PVIIR(i-1,j)+PVIIR(i-1,j+1)+PVIIR(i-1,j+2)+PVIIR(i,j-2)+PVIIR(i,j-1)+PVORI(i,j))/13 (1)
equation (1) the target pixel P (i, j) becomes an IIR pixel after the calculation is completed, and becomes a reference pixel in the next round of IIR filtering (with the pixel P (i, j +1) of fig. 1 as the target pixel). Because the pixel referred by the IIR filtering is an IIR pixel instead of an original pixel, and the IIR pixel is less interfered by noise compared with the original pixel, the IIR filtering can use less reference pixels (meaning that hardware resources can be saved) to carry out effective low-pass filtering processing so as to remove high-frequency noise in the image. However, in the top-left-to-bottom-right scanning order as an example, the IIR filter only takes the left, top-left, top, and/or top-right pixels of the target pixel as a reference, and does not consider the pixels in other directions, so that an image dragging (image dragging) phenomenon in which the left, top-left, top, and/or top-right pixel features spread in other directions is likely to occur. This image dragging phenomenon is prone to cause color distortion, and if it occurs at the edge of the image, it will cause the image edge to be blurred. Therefore, how to effectively remove the high frequency noise and avoid the image dragging phenomenon and color distortion is an important issue in the field.
Disclosure of Invention
In view of the foregoing, it is an object of the present invention to provide an image processing method and an image processing apparatus.
The invention discloses an image processing method for filtering an image, which comprises the following steps: (A) generating a first weight for each first reference pixel, wherein the image comprises a plurality of first reference pixels, and the magnitude of the first weight is related to the similarity between a target pixel and the first reference pixel corresponding to the first weight; (B) calculating a first filtering value of the target pixel according to the first weights, the pixel value of the target pixel and the pixel values of the first reference pixels; (C) generating a second weight for each second reference pixel, wherein the image comprises a plurality of second reference pixels, and the magnitude of the second weight is related to the similarity between the target pixel and the second reference pixel corresponding to the second weight; and (D) calculating a second filtered value of the target pixel according to the second weights, the first filtered value of the target pixel and the pixel values of the second reference pixels.
The invention further discloses an image processing apparatus for filtering an image, comprising a first weight calculating circuit, a first filtering value calculating circuit, a second weight calculating circuit and a second filtering value calculating circuit. The first weight calculation circuit generates a first weight for each first reference pixel, wherein the image comprises a plurality of first reference pixels, and the magnitude of the first weight is related to the similarity between a target pixel and the first reference pixel corresponding to the first weight. The first filtering value calculating circuit calculates a first filtering value of the target pixel according to the first weights, the pixel value of the target pixel and the pixel values of the first reference pixels. The second weight calculation circuit generates a second weight for each second reference pixel, wherein the image comprises a plurality of second reference pixels, and the magnitude of the second weight is related to the similarity between the target pixel and the second reference pixel corresponding to the second weight. The second filtering value calculating circuit calculates a second filtering value of the target pixel according to the second weights, the first filtering value of the target pixel and the pixel values of the second reference pixels.
The present invention further discloses an image processing method for filtering an image, comprising the following steps: (A) generating a weight for each reference pixel, wherein the image comprises a plurality of reference pixels, and the size of the weight is related to the similarity between a target pixel and the reference pixel corresponding to the weight; and (B) performing an infinite impulse response filtering operation according to the weights, the pixel value of the target pixel and the pixel values of the reference pixels to obtain an infinite impulse response filtering value of the target pixel.
The image processing method and the image processing device of the invention implement an edge protection mechanism to inhibit the image dragging phenomenon of IIR filtering. Compared with the prior art, the invention can protect the edge structure of the image while filtering.
The features, implementations and functions of the present invention will be described in detail with reference to the drawings.
Drawings
FIG. 1 is a schematic view of a window for image processing;
FIG. 2 is a functional block diagram of an image processing apparatus according to an embodiment of the present invention;
FIGS. 3A-3B are flowcharts of an image processing method according to an embodiment of the present invention;
FIG. 4 is a schematic view of another window for image processing;
FIG. 5 is a block diagram of an embodiment of a weight calculation circuit;
FIGS. 6A-6B are diagrams showing an example of a relationship between weights and pixel differences;
FIG. 7 is a block diagram of an image processing apparatus according to another embodiment of the present invention;
FIG. 8 is a flowchart illustrating an image processing method according to another embodiment of the present invention; and
FIG. 9 is a schematic view of another window for image processing.
[ notation ] to show
200 image processing device
210. 710 memory
220. 720, 740 weight calculation circuit
230. 730, 750 filter circuit
510 pixel similarity calculation circuit
520 weight determination circuit
S310 to S340, S810 to S860 steps
Detailed Description
The technical terms in the following description refer to the conventional terms in the technical field, and some terms are explained or defined in the specification, and the explanation of the some terms is based on the explanation or the definition in the specification.
The disclosure of the present invention includes an image processing method and an image processing apparatus. Since some of the components included in the image processing apparatus of the present invention may be known components alone, the following description will omit details of the known components without affecting the full disclosure and feasibility of the present invention.
Fig. 2 is a functional block diagram of an image processing apparatus according to an embodiment of the present invention, and fig. 3 is a flowchart corresponding to the image processing method of fig. 2. The image processing apparatus 200 includes a memory 210, a weight calculating circuit 220, and a filtering circuit 230. When the image processing apparatus 200 performs image processing, a target pixel P (i, j) is obtained, and a plurality of pixels adjacent to the target pixel P (i, j) are used as reference pixels (step S310), wherein the reference pixels are IIR pixels. Taking fig. 1 as an example, the reference pixels may include the 12 IIR pixels, and the memory 210 needs to store at least the 12 IIR pixels. FIG. 4 shows another embodiment of a window used by the image processing apparatus 200. In this case, the reference pixel only includes two IIR pixels, and the memory 210 needs to store at least two IIR pixels. Compared to fig. 1, the embodiment of fig. 4 can store at least 10 less pixel values of the reference pixels.
Next, the weight calculating circuit 220 generates a weight for each reference pixel (step S320), in detail, the weight calculating circuit 220 obtains the reference pixel from the memory 210, and determines a weight W for each reference pixel according to the similarity between the reference pixel and the target pixel, i.e. the weight W is related to the similarity between two pixels. Fig. 5 is a functional block diagram of an embodiment of the weight calculating circuit 220. First, the pixel similarity calculation circuit 510 calculates the pixel value PV of the target pixel tarAnd a pixel value PV of the reference pixelrefThe similarity between each reference pixel and the target pixel is calculated to obtain a difference d (step S322). For example, the pixel similarity calculation circuit 510 calculates the absolute difference d ═ PV of the twotar-PVrefIf the absolute difference d is larger, the similarity between the target pixel and the reference pixel is lower. Therein, theIn the examples, the pixel value PVtarIs the original pixel value PV of the target pixel P (i, j)ORI(i, j), pixel value PVrefIIR filtered value PV for reference pixelIIR。
The weight determination circuit 520 then determines a weight W for each reference pixel according to the difference d (equivalently, according to the similarity of the pixels), and the maximum weight Wmax, the first threshold th1 and the first threshold th2 (step S324). Fig. 6A and 6B are diagrams showing an example of a relationship between the weight W and the difference d. When the difference d is smaller than the first threshold th1, which indicates that the similarity between the target pixel and the reference pixel is high, the weight determination circuit 520 gives the highest weight Wmax to the reference pixel; when the difference d is between the first threshold th1 and the second threshold th2, the weight W decreases as the difference d increases; when the difference d is greater than the second threshold th2, which indicates that the similarity between the target pixel and the reference pixel is low, the weight W given to the reference pixel by the weight determination circuit 520 is zero. For hardware implementation convenience, the weight determining circuit 520 may automatically divide the average value between the first threshold th1 and the first threshold th2 into q equal parts (q is an integer greater than 1, and q is 3 in the example of fig. 6B) so that the weight W is changed in a stepwise manner (stepwise). Compared to the implementation of fig. 6A, fig. 6B may omit the divider for calculating the slope between the threshold values th1 and th2 of fig. 6A.
After obtaining the weight of each reference pixel, the filter circuit 230 calculates a weighted average of the target pixel and the reference pixel according to equation (2) to obtain an IIR filtered value of the target pixel (step S330).
PVIIR(i，j)＝((WSUM-WNBR)×PVORI(i，j)+∑(W×PVIIR))/WSUM (2)
As shown in FIG. 3B, step S330 may be subdivided into calculating a plurality of first products (WXPV)IIR) (step S332), a second product ((W) is calculatedSUM-WNBR)×PVORI(i, j)) (step S334), and obtaining a filtered value of the target pixel according to the first products and the second products (step S336). In detail, in step S336, the sum of the first products and the second products is divided by the user definitionIs weighted sum W ofSUMTo obtain a filtered value of the target pixel. Wherein, WNBRThe first product (W × PV) being the sum of the weights of all reference pixelsIIR) Is the product of the weight W of a reference pixel and the IIR filtered value of the reference pixel, Σ (W × PV)IIR) Is the sum of the first products of all reference pixels. For example, taking fig. 4 as an example, equation (2) becomes:
PVIIR(i，j)＝((WSUM-WNBR)×PVORI(i，j)+W(i-1，j)×PVIIR(i-1，j)+W(i，j-1)×PVIIR(i，j-1))/WSUM (3)
wherein the content of the first and second substances,
WNBR＝W(i-1，j)+W(i，j-1) (4)
the sum of the weights can be designed as WSUM- (n-1) × Wmax ≈ Wmax, n is the number of pixels included in the window. That is, when the similarity between all the reference pixels and the target pixel in the window is very high (e.g., the window is located in a flat area where the pixel value in the image does not change much), the weight (W) of the target pixel SUM-WNBR＝WSUM- (n-1) × Wmax ≈ Wmax) is comparable to the weight of any reference number of phases. In one embodiment, WSUMThe power of 2 can be designed, so that the division of equation 2 can be implemented by shifting the value of the binary without using a divider.
When the target pixel is located on the edge of the image, the reference pixel in the window that is also located on the edge (i.e. the reference pixel with high similarity to the target pixel) has a higher weight W, while the other reference pixels that are not located on the edge (i.e. the reference pixel with low similarity to the target pixel) have a lower weight W, so that the edge feature of the filtered image is retained and the weight can be considered as an edge preserving (edge preserving) weight. In detail, the mechanism for generating the weight W by the weight calculation circuit 220 is related to the edge of the image, and this mechanism enables the filtering process of the image processing apparatus 200 to have the edge perception capability, so that the IIR filtering image dragging phenomenon can be effectively suppressed to prevent the color distortion and the edge blur.
Obtaining an IIR filtering value PV of the target pixelIIRThen, the image processing apparatus 200 outputs the data to a circuit (not shown) at a later stage (step S340), and stores the data in the memory 210 for use in a subsequent IIR filtering process.
Fig. 7 is a functional block diagram of an image processing apparatus according to another embodiment of the present invention, and fig. 8 is a flowchart corresponding to the image processing method of fig. 7. The image processing apparatus 700 includes a memory 710, weight calculation circuits 720 and 740, and filter circuits 730 and 750. The memory 710 stores reference pixels. When the image processing apparatus 700 performs image processing, a target pixel P (i, j) is obtained, and pixels adjacent to the target pixel P (i, j) are used as reference pixels (step S810). Then, the weight calculation circuit 720 uses the original pixel values PV of the target pixel and the reference pixelORIA first weight W1 is generated for each reference pixel (step S820), i.e., the first weight W1 is related to the similarity between two pixels. Similarly, a functional block diagram of an embodiment of the weight calculating circuit 720 is shown in fig. 5, and the step S820 includes two sub-steps S322 and S324. In step S820, the pixel value PV of the pixel similarity calculation circuit 510 is inputtarAnd pixel value PVrefAre the original pixel values PV of the target pixels P (i, j), respectivelyORIAnd an original pixel value PV of the reference pixelORI。
The filter circuit 730 then calculates the weighted average of the target pixel and the reference pixel according to equation (3) to obtain the edge-protection low-pass filter value PV of the target pixel EPF(i, j) (step S830).
PVEPP(i，j)＝((WSUM-WNBR)×PVORI(i，j)+∑(w1×PVORI))/WSUM (5)
Equation (5) is a plurality of first products (W1 XPV)ORI) With the second product ((W)SUM-WNBR)×PVORI(i, j)) divided by a user-defined weighted sum WSUMThe result of (1). Wherein, W1NBRThe sum of the first weights W1 for all reference pixels, the first product (W1 XPV)ORI) A first weight W1 of a reference pixel and the reference pixelIs multiplied by the original pixel value of (W1 × PV)ORI) Is the sum of the first products of all reference pixels. For example, taking the window of fig. 9 as an example, equation (5) becomes:
PVEPF(i，j)＝((WSUM-W1NBR)×PVORI(i，j)+W1(i-1，j-1)×PVORI(i-1，j-1)+W1(i-1，j)×PvORI(i-1，j)+W1(i-1，j+1)×PVORI(i-1，j+1)+W1(i，j-1)×PVORI(i，j-1)+W1(i，j+1)×PVORI(i，j+1)+W1(i+1，j-1)×PVORI(i+1，j-1)+W1(i+1，j)×PVORI(i+1，j)+W1(i+1，j+1)×PVORI(i+1，j+1))/WSUM (6)
wherein the content of the first and second substances,
W1NBR＝W1(i-1，j-1)+W1(i-1，j)+W1(i-1，j+1)+W1(i，j-1)+W1(i，j+1)+W1(i+1，j-1)+W1(i+1，j)+W1(i+1，j+1) (7)
in fact, steps S820-S830 are Finite Impulse Response (FIR) filtering processes with edge protection mechanisms. Edge-protected low-pass filtered value PV of target pixelEPF(i, j) is stored in the memory 710 for later filtering.
Next, in steps S840 and S850, the weight calculation circuit 740 and the filter circuit 750 apply the edge protection low pass filter PVEPF(i, j) further performing IIR filtering processing to obtain an IIR filtering value PV of the target pixelIIR(i, j). That is, steps S840 and S850 are substantially the same as steps S320 and S330, except that the weight calculation circuit 740 and the filter circuit 750 use the edge-protected low-pass filter PV of the target pixel EPF(i, j) instead of the original pixel value PVORI (I, j) similarly, a functional block diagram of an embodiment of the weight calculating circuit 740 is shown in fig. 5, and in step S840, the pixel value PV of the pixel similarity calculating circuit 510 is inputtedtarAnd pixel value PVrefAre edge-protected low-pass filtered values PV of the target pixels P (i, j), respectivelyEPF(i, j) and IIR filtered value PV of reference pixelIIR. After obtaining a second weight W2 for each reference pixel (similarly, the weight W2 is related to the similarity between two pixels), the filter circuit 750 calculates the IIR filter value of the target pixel according to equation (8).
PVIIR(i，j)＝((WSUM-W2XBR)×PVEPF(i，j)+∑(W2×PVIIR)/WSUM (8)
Equation (8) is similar to equation (2) and will not be described again. Wherein, W2NBRIs the sum of the second weights W2 for all reference pixels, taking figure 4 as an example,
W2NBR＝W2(i-1，j)+W2(i，j-1) (9)
the embodiment of fig. 7 and 8 performs the filtering process twice on the image, the first time being performed by the weight calculation circuit 720 and the filter circuit 730 (corresponding to steps S820 to S830), and the second time being performed by the weight calculation circuit 740 and the filter circuit 750 (corresponding to steps S840 to S850). The mechanism of the second filtering process is similar to that of the filtering process of the previous embodiment (corresponding to fig. 2 and fig. 3), and the purpose of the first filtering process is to further remove the high-frequency noise of the image, so as to make the image more clear. The two filtering processes of the embodiments of fig. 7 and fig. 8 implement the edge protection mechanism, i.e. the first weight and the second weight are both edge protection weights, so the obtained filtering result can effectively suppress the image dragging phenomenon to prevent the color distortion and the edge blur. The reference pixels of the first filtering process of the embodiments of fig. 7 and 8 include the reference pixels of the second filtering process; for example, as shown in FIGS. 9 and 4, the two reference pixels of FIG. 4 (P (i-1, j) and P (i, j-1)) are part of the eight reference pixels of FIG. 9.
The invention determines the allowable range of pixel similarity by two threshold values, so that a better trade-off (trade-off) can be easily and effectively obtained between the suppression degree of the image dragging phenomenon and the noise removal degree. In particular, when the first and second thresholds are simultaneously adjusted downward, the reference pixel must be closer to the target pixel to be given the highest weight Wmax, thereby helping to suppress the image dragging phenomenon (i.e., enhance the edge protection of the image); when the first and second thresholds are simultaneously raised, the reference pixel has a greater probability of being given the highest weight Wmax, thereby facilitating noise removal (i.e., making the image sharper). Increasing the difference between the first and second thresholds provides better and more adaptive operation flexibility in the filtering process between suppressing the image dragging phenomenon and removing the noise.
The weight calculation circuit and the filter circuit of the invention can be combined by an adder, a subtracter, a multiplier, a divider, a comparator, a multiplexer, a logic circuit and other elements. The weight calculation circuit 720 and the weight calculation circuit 740 in FIG. 7 can be the same hardware circuit or separate hardware circuits, and the filter circuit 730 and the filter circuit 750 can be the same hardware circuit or separate hardware circuits. The image processing device and the image processing method are suitable for images with various color formats (such as RGB channels, YUV channels, gray-scale images and the like), and each channel of the color formats is independently processed, so that an edge protection mechanism (namely, the implementation of edge protection weight) is beneficial to avoiding the occurrence of color distortion.
Since the details and variations of the present invention can be understood by those skilled in the art from the disclosure of the present invention, the repetitive description is omitted here for the sake of avoiding unnecessary detail without affecting the disclosure requirements and the feasibility of the present invention. It should be noted that the shapes, sizes, proportions, and sequence of steps of the elements and steps shown in the drawings are illustrative only and are not intended to be limiting, since those skilled in the art will understand the present invention.
Although the embodiments of the present invention have been described above, these embodiments are not intended to limit the present invention, and those skilled in the art can make variations on the technical features of the present invention according to the explicit or implicit contents of the present invention, and all such variations may fall within the scope of the patent protection sought by the present invention.

Claims (10)

1. An image processing method for filtering an image, comprising:
(A) generating a first weight for each first reference pixel, wherein the image comprises a plurality of first reference pixels, and the magnitude of the first weight is related to the similarity between a target pixel and the first reference pixel corresponding to the first weight;
(B) Calculating a first filtering value of the target pixel according to the first weights, the pixel value of the target pixel and the pixel values of the first reference pixels;
(C) generating a second weight for each second reference pixel, wherein the image includes a plurality of second reference pixels, and the magnitude of the second weight is related to the similarity between the target pixel and the second reference pixel corresponding to the second weight, wherein the higher the similarity between the target pixel and the second reference pixel corresponding to the second weight is, the larger the second weight is; and
(D) and calculating a second filter value of the target pixel according to the second weights, the first filter value of the target pixel and the pixel values of the second reference pixels, wherein the pixel values of the second reference pixels are infinite impulse response filter values of the second reference pixels, and the second filter value of the target pixel is infinite impulse response filter value of the target pixel.
2. The image processing method as claimed in claim 1, wherein the step (B) comprises:
calculating a plurality of first products, wherein the first products are the products of the pixel values of the first reference pixels and the corresponding first weights;
Calculating a second product, wherein the second product is the product obtained by multiplying the pixel value of the target pixel by a difference value of the sum of a preset weight and the first weights; and
obtaining the first filtering value of the target pixel according to the first products and the second products;
and the step (D) comprises:
calculating a plurality of third products, wherein the third products are the products of the pixel values of the second reference pixels and the corresponding second weights;
calculating a fourth product, wherein the fourth product is a product obtained by multiplying the first filtering value of the target pixel by a difference value of the sum of the preset weight and the second weights; and
and obtaining the second filtering value of the target pixel according to the third products and the fourth product.
3. The image processing method of claim 1, wherein the pixel value of the first reference pixels is an original pixel value of the first reference pixels, and the pixel value of the second reference pixels is the second filtered value of the second reference pixels.
4. The method of claim 1, wherein the steps (A) and (B) are performed by calculating original pixel values of the first reference pixels and the target pixel, and the steps (C) and (D) are performed by calculating the second filtered values of the second reference pixels and the first filtered value of the target pixel.
5. The image processing method as claimed in claim 4, wherein the step (A) or the step (C) comprises:
calculating an absolute difference value between the first reference pixel or the second reference pixel and the target pixel; and
determining the first weight or the second weight according to the absolute difference, a first threshold and a second threshold.
6. The image processing method as claimed in claim 1, wherein the second reference pixels are a portion of the first reference pixels.
7. An image processing apparatus for filtering an image, comprising:
a first weight calculating circuit, generating a first weight for each first reference pixel, wherein the image comprises a plurality of first reference pixels, and the magnitude of the first weight is related to the similarity between a target pixel and the first reference pixel corresponding to the first weight;
a first filter value calculating circuit for calculating a first filter value of the target pixel according to the first weights, the pixel value of the target pixel and the pixel values of the first reference pixels;
a second weight calculating circuit, configured to generate a second weight for each second reference pixel, wherein the image includes a plurality of second reference pixels, and the magnitude of the second weight is related to the similarity between the target pixel and the second reference pixel corresponding to the second weight, and the higher the similarity between the target pixel and the second reference pixel corresponding to the second weight is, the larger the second weight is; and
And a second filter value calculation circuit for calculating a second filter value of the target pixel according to the second weights, the first filter value of the target pixel and the pixel values of the second reference pixels, wherein the pixel values of the second reference pixels are the infinite impulse response filter values of the second reference pixels, and the second filter value of the target pixel is the infinite impulse response filter value of the target pixel.
8. The image processing device as claimed in claim 7, wherein the pixel value of the first reference pixels is an original pixel value of the first reference pixels, and the pixel value of the second reference pixels is the second filtered value of the second reference pixels.
9. An image processing method for filtering an image, comprising:
(A) generating a weight for each reference pixel, wherein the image comprises a plurality of reference pixels, and the size of the weight is related to the similarity between a target pixel and the reference pixel corresponding to the weight, wherein the higher the similarity between the target pixel and the reference pixel corresponding to the weight is, the larger the weight is; and
(B) and performing an infinite impulse response filtering operation according to the weights, the pixel value of the target pixel and the pixel values of the reference pixels to obtain an infinite impulse response filtering value of the target pixel, wherein the pixel values of the reference pixels are the infinite impulse response filtering values of the reference pixels.
10. The image processing method as claimed in claim 9, wherein the step (B) comprises:
calculating a plurality of first products, wherein the first products are the products of the pixel values of the reference pixels and the corresponding weights;
calculating a second product, wherein the second product is the product obtained by multiplying the pixel value of the target pixel by a difference value of the sum of a preset weight and the weights; and
and obtaining the infinite impulse response filter value of the target pixel according to the first products and the second products.
CN201810062839.6A 2018-01-23 2018-01-23 Image processing method and image processing device Active CN110072034B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810062839.6A CN110072034B (en) 2018-01-23 2018-01-23 Image processing method and image processing device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810062839.6A CN110072034B (en) 2018-01-23 2018-01-23 Image processing method and image processing device

Publications (2)

Publication Number Publication Date
CN110072034A CN110072034A (en) 2019-07-30
CN110072034B true CN110072034B (en) 2021-10-08

Family

ID=67364818

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810062839.6A Active CN110072034B (en) 2018-01-23 2018-01-23 Image processing method and image processing device

Country Status (1)

Country Link
CN (1) CN110072034B (en)

Family Cites Families (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4086479B2 (en) * 2001-03-23 2008-05-14 Ｎｅｃディスプレイソリューションズ株式会社 Image quality improving apparatus and image quality improving method
US7822285B2 (en) * 2004-05-20 2010-10-26 Omnivision Technologies, Inc. Methods and systems for locally adaptive image processing filters
US7769089B1 (en) * 2004-12-02 2010-08-03 Kolorific, Inc. Method and system for reducing noise level in a video signal
CN100559836C (en) * 2007-09-20 2009-11-11 宝利微电子系统控股公司 A kind of digital image noise reduction method
KR101517233B1 (en) * 2008-08-12 2015-05-15 엘지전자 주식회사 Apparatus for removing noise using moving estimation
JP2013059016A (en) * 2011-08-12 2013-03-28 Sony Corp Image processing device, method, and program
CA2861220C (en) * 2012-01-24 2020-02-11 Sony Corporation Image processing apparatus and image processing method, and program
CN103581507A (en) * 2013-08-21 2014-02-12 成都云鹰科技有限公司 Method dynamically adjusting threshold value through mean square error in de-noising algorithm
KR102105102B1 (en) * 2013-10-10 2020-04-27 삼성전자주식회사 Display device and method thereof
CN103871031A (en) * 2014-03-07 2014-06-18 西安电子科技大学 Kernel regression-based SAR image coherent speckle restraining method
CN107317959B (en) * 2016-04-26 2020-01-14 瑞昱半导体股份有限公司 Image filtering device and image filtering method thereof

Also Published As

Publication number Publication date
CN110072034A (en) 2019-07-30

Similar Documents

Publication Publication Date Title
JP5979856B2 (en) Image blur detector
JP4290193B2 (en) Image processing device
KR100532121B1 (en) Method and apparatus for image detail enhancement
US8189947B2 (en) Image edge correction apparatus and method
TWI463861B (en) Device and method for removing mosquito noises
WO2011033619A1 (en) Image processing device, image processing method, image processing program, and storage medium
CN109389560B (en) Adaptive weighted filtering image noise reduction method and device and image processing equipment
WO2004002135A1 (en) Motion detection device and noise reduction device using that
JP5105286B2 (en) Image restoration apparatus, image restoration method, and image restoration program
JP3240371B2 (en) Video signal processing device
JP2011065339A (en) Image processing apparatus, image processing method, image processing program and storage medium
JP7265316B2 (en) Image processing device and image processing method
JP2007306505A (en) Apparatus and method for improving picture quality
JP6242157B2 (en) Image processing apparatus, image processing method, and program
CN110072034B (en) Image processing method and image processing device
JP4333150B2 (en) Signal processing apparatus and method, recording medium, and program
TWI712989B (en) Image processing method and image processing device
JP2007265122A (en) Image processor
JP6644877B2 (en) Image processing device
JP2003087605A (en) Device and method for reducing noise, medium, and program
JP3106831B2 (en) Video signal processing device
JP2011217001A (en) Filter device, and filtering method
KR100821939B1 (en) Apparatus and method for image noise reduction
JP2010171624A (en) Outline correction circuit and interpolation pixel generation circuit
JP2008021219A (en) Image processing apparatus and method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant