An image detail enhancement method, apparatus, computer device, and storage medium

Abstract

This invention relates to the field of image processing and discloses an image detail enhancement method, apparatus, computer device, and storage medium. The method provided by this invention includes: performing a first color space conversion operation on an input image and extracting the luminance component image L and other component images corresponding to the converted color space; performing a low-pass filter on the luminance component image L to obtain a filtered result image Lg; performing a differential transformation on the filtered result image Lg using a first preset transformation formula to obtain a difference result image Ld; performing a nonlinear transformation on the difference result image Ld using a second preset transformation formula to obtain a nonlinear transformed image Lt; fusing the nonlinear transformed image Lt with other component images, and performing a second color space conversion operation on the fused image to obtain an output image. Compared with existing technologies, the method provided by this invention has lower overall algorithm complexity, is easier to implement in software, and provides superior detail enhancement effects.

Description

Technical Field

[0001] This invention relates to the field of image processing, and more specifically to an image detail enhancement method, apparatus, computer device, and storage medium. Background Technology

[0002] The texture, contours, and other details of an image are of great value in practical applications, and the quality of the image directly affects the results. While existing image detail enhancement methods, such as multi-scale image detail enhancement, multi-exposure fusion algorithms, and neural network-based image detail enhancement algorithms, can achieve certain enhancement effects, each has its drawbacks. For example, multi-scale image detail enhancement can improve image detail, but it can cause color casts in color images; multi-exposure fusion algorithms can suffer from slight halo effects, leading to poor enhancement results; and neural network-based image detail enhancement algorithms are highly complex and not easily applied in real-time to specific image processing systems. Summary of the Invention

[0003] In view of this, the present invention provides an image detail enhancement method, apparatus, computer device and storage medium to overcome the shortcomings of existing image detail enhancement algorithms, such as poor enhancement effect and high algorithm complexity that makes them difficult to implement in software.

[0004] In a first aspect, the present invention provides an image detail enhancement method, comprising:

[0005] Perform a first color space conversion operation on the input image, and extract the luminance component image L and other component images corresponding to the converted color space;

[0006] The brightness component image L is low-pass filtered to obtain the filtered result image Lg;

[0007] The filtered result image Lg is subjected to a differential transformation using the first preset transformation formula to obtain the difference result image Ld;

[0008] The difference result image Ld is subjected to a nonlinear transformation using a second preset transformation formula to obtain a nonlinear transformed image Lt;

[0009] The nonlinear transformed image Lt is fused with the other component images, and the fused image is then subjected to a second color space conversion operation to obtain the output image.

[0010] The image detail enhancement method provided in this embodiment has lower overall algorithm complexity, is easier to implement in software, and has better detail enhancement effect compared with the prior art.

[0011] In one optional implementation, the process of performing color space conversion on the image to be processed to extract the luminance component image L and other component images includes:

[0012] The other component images include color dimension component image a and color dimension component image b;

[0013] The step of performing a first color space conversion operation on the input image and extracting the luminance component image L and other component images corresponding to the converted color space includes:

[0014] The input image is subjected to pixel value normalization processing, wherein the input image is in RGB color space, and the pixel value range of the input image after normalization processing is [0, 1].

[0015] The input image after normalization is converted from the RGB color space to the Lab color space;

[0016] Extract the luminance component image L, color dimension component image a, and color dimension component image b corresponding to the Lab color space.

[0017] In this embodiment of the invention, the image to be processed is converted from the RGB color space to the Lab color space. The luminance component image L it includes closely matches human luminance perception. Therefore, extracting the L component facilitates subsequent steps in extracting non-detail and detail information from the luminance component image.

[0018] In one optional implementation, the step of fusing the nonlinear transformed image Lt with the other component images, and performing a second color space conversion operation on the fused image to obtain an output image, includes:

[0019] The nonlinear transformed image Lt is fused with the color dimension component image a and the color dimension component image b;

[0020] The fused image is converted from the Lab color space to the RGB color space, and the converted image is then subjected to pixel value conversion to obtain an output image with a pixel value range of [0, 255].

[0021] The present invention fuses the color dimension component images a and b, which have a bright effect after processing the luminance component, to obtain an image with good image detail enhancement.

[0022] In one optional implementation, the process of performing low-pass filtering on the luminance component image L to obtain the filtered result image Lg includes:

[0023] Determine the target pixel block corresponding to each pixel in the luminance component image L, wherein the size of the target pixel block is the same as the size of the preset Gaussian low-pass filter kernel template, and the radius of the preset Gaussian low-pass filter kernel template is 1 / 4*min(m,n), where m and n are the length and width of the luminance component image L, respectively, and min(m,n) represents the minimum value of m and n.

[0024] For each target pixel block, the template is used as a weight and multiplied by the pixel value of each target pixel in the target pixel block, and the multiplied values ​​are summed to obtain a replacement pixel value. The pixel value of the center point of the pixel block is then updated using the replacement pixel value.

[0025] After each pixel in the luminance component image L is updated, the image composed of all the updated pixels is determined as the filtered result image Lg.

[0026] In this embodiment of the invention, the luminance component image L is subjected to Gaussian low-pass filtering to obtain the filtered result image Lg, with the aim of extracting non-detail information from the luminance component image.

[0027] In one optional implementation, the first preset transformation formula is:

[0028] Ld = Lg - L.

[0029] The purpose of using the transformation formula to perform differential transformation in this embodiment of the invention is to extract image detail information of the luminance component.

[0030] In one optional implementation, the second preset transformation formula is:

[0031] Lt=L-β*Ld α

[0032] Where α is the exponential coefficient and β is the linear coefficient, and their values ​​satisfy 0 < α < 1 and 0 < β < 1.

[0033] The purpose of using the transformation formula to perform nonlinear transformation on the image in this embodiment of the invention is to enhance the details in the luminance component image.

[0034] In a second aspect, embodiments of the present invention provide an image detail enhancement apparatus, the apparatus comprising:

[0035] The image luminance component extraction module is used to perform a first color space conversion operation on the input image and extract the luminance component image L and other component images corresponding to the converted color space.

[0036] The low-pass filter module is used to perform low-pass filtering on the luminance component image L to obtain the filtered result image Lg;

[0037] The differential transformation module is used to perform differential transformation on the filtered result image Lg using a first preset transformation formula to obtain the difference result image Ld.

[0038] The nonlinear transformation module is used to perform a nonlinear transformation on the difference result image Ld using a second preset transformation formula to obtain a nonlinear transformed image Lt.

[0039] An enhanced image output module is used to fuse the nonlinear transformed image Lt with the other component images and convert it to the original image color space to obtain an output image.

[0040] In one optional implementation, the image luminance component extraction module includes:

[0041] The normalization unit is used to normalize the pixel values ​​of the input image, wherein the input image is in RGB color space and the pixel values ​​of the input image after normalization are in the range of [0, 1].

[0042] The first space conversion unit is used to convert the normalized input image from the RGB color space to the Lab color space.

[0043] The component extraction unit is used to extract the luminance component image L, color dimension component image a, and color dimension component image b corresponding to the Lab color space.

[0044] Thirdly, embodiments of the present invention provide a computer device, including:

[0045] A memory and a processor are communicatively connected, the memory storing computer instructions, and the processor executing the computer instructions to perform the image detail enhancement method provided in any embodiment of the first aspect.

[0046] Fourthly, according to an embodiment of the present invention, a computer-readable storage medium is provided, wherein computer instructions are stored on the computer-readable storage medium, the computer instructions being configured to cause a computer to perform the image detail enhancement method provided in any embodiment of the first aspect of the invention. Attached Figure Description

[0047] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0048] Figure 1 This is a flowchart illustrating an image detail enhancement method according to an embodiment of the present invention;

[0049] Figure 2 (a) and (b) Comparison of images before and after processing in embodiments of the present invention;

[0050] Figure 3 This is a functional block diagram of an image detail enhancement device according to an embodiment of the present invention;

[0051] Figure 4 This is a functional block diagram of a specific example of an image detail enhancement apparatus according to an embodiment of the present invention;

[0052] Figure 5 This is a functional block diagram of a computer device provided according to an embodiment of the present invention. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] This embodiment provides an image detail enhancement method, such as Figure 1 As shown, the process includes the following steps:

[0055] Step S101: Perform a first color space conversion operation on the input image, and extract the luminance component image L and other component images corresponding to the converted color space.

[0056] In this embodiment of the invention, the image to be processed is an RGB format image with a bit width of 8. RGB image data has pixel values ​​that are integers between 0 and 255. To reduce the computational load, the pixel values ​​are normalized, typically by dividing the pixel value by 255 to obtain a value between 0 and 1 for calculation. Further, the normalized image is converted from the RGB color space to the Lab color space, and the luminance component image L, color dimension component image a, and color dimension component image b are extracted. The purpose of this color space conversion is to extract the luminance component of the image, providing image data for subsequent image detail enhancement. The biggest advantage of the RGB color space is its suitability for display systems and its intuitiveness. However, the three color components are highly correlated; any change in luminance will cause a corresponding change in all three components. If a certain component of a color changes to a certain extent, the overall color may change. Therefore, other color spaces with luminance components are needed.

[0057] This invention converts the image to be processed from the RGB color space to the Lab color space. The Lab color space is a CIE XYZ color space based on non-linear compression. Its coordinates are a color-opposite space, with dimension L representing brightness, and a and b representing color opposites. Color a includes a range from dark green (low brightness value) to gray (medium brightness value) to bright pink (high brightness value); color b includes a range from bright blue (low brightness value) to gray (medium brightness value) to yellow (high brightness value). Therefore, this color mixing will produce a bright effect. Its L component closely matches human brightness perception, so extracting the L component facilitates subsequent steps in extracting non-detail and detail information from the brightness component image. In other embodiments, it can be converted to other color spaces with brightness components; however, the Lab color space conversion in this embodiment yields the best results.

[0058] Step S102: Perform low-pass filtering on the luminance component image L to obtain the filtered result image Lg.

[0059] This invention employs Gaussian low-pass filtering on the luminance component image L to obtain the filtered result image Lg. The aim is to extract non-detail information from the luminance component image. Gaussian low-pass filtering is the most effective and easiest method to implement. The specific implementation process is as follows:

[0060] Each pixel in the luminance component image L corresponds to a target pixel block, the size of which is the same as the size of a preset Gaussian low-pass filter kernel template. The radius of the preset Gaussian low-pass filter kernel template is 1 / 4*min(m,n), where m and n are the length and width of the luminance component image L, respectively, and min(m,n) represents the minimum value of m and n. For each target pixel block, the template is multiplied by the pixel value of each target pixel in the block as a weight, and the results are summed to obtain a replacement pixel value. This replacement pixel value is then used to update the pixel value of the center point of the pixel block. After updating each pixel in the luminance component image L, the image formed by all the updated pixels is determined as the filtered result image Lg. It should be noted that when the center point of a target pixel block is located at the edge or corner of the luminance component image L, its missing neighboring pixels are replaced with the pixel value of the center point itself.

[0061] Step S103: Perform differential transformation on the filtered result image Lg using the first preset transformation formula to obtain the difference result image Ld.

[0062] The purpose of the differential transformation in this embodiment of the invention is to extract detail information from the luminance component image. Specifically, the filtered image Lg is subjected to differential transformation using a first preset transformation formula to obtain the difference result image Ld. The first preset transformation formula is:

[0063] Ld = Lg - L.

[0064] Step S104: Perform a nonlinear transformation on the difference result image Ld using the second preset transformation formula to obtain the nonlinear transformed image Lt.

[0065] In this embodiment of the invention, the nonlinear transformation of the image enhances the details in the luminance component image. Specifically, the difference result image Ld is nonlinearly transformed using the following second preset transformation formula to obtain the nonlinearly transformed image Lt. The second preset transformation formula is:

[0066] Lt=L-β*Ld α

[0067] Where α is the exponential coefficient and β is the linear coefficient, and their values ​​satisfy 0 < α < 1 and 0 < β < 1. The specific values ​​are determined based on actual experience or processing results.

[0068] Step S105: The nonlinear transformed image Lt is fused with other component images, and the fused image is subjected to a second color space conversion operation to obtain the output image.

[0069] Specifically, in this embodiment of the invention, the nonlinear transformed image Lt is fused with color dimension component images a and b; the fused image is converted from the Lab color space to the RGB color space. At this time, the pixel value range after the space conversion is still [0, 1]. It is necessary to further convert the pixel value of the space-converted image, that is, multiply the pixel value by 255 to obtain an output image with a pixel value range of [0, 255], which is used as the image after the original image enhancement processing.

[0070] In one embodiment, before processing, as follows Figure 2 As shown in (a), the processed image is as follows: Figure 2 As shown in (b), by magnifying a portion of the stone, it can be seen that after processing by the image detail enhancement method provided in this embodiment of the invention, the texture and layering details of the stone are enhanced, making the image information clearer. Compared with existing technologies, the image detail enhancement method provided in this embodiment has lower overall algorithm complexity, is easier to implement in software, and achieves better processing results.

[0071] This embodiment provides an image detail enhancement device, such as... Figure 3 As shown, it includes:

[0072] The image luminance component extraction module 301 is used to perform a first color space conversion operation on the input image and extract the luminance component image L and other component images corresponding to the converted color space.

[0073] The low-pass filtering module 302 is used to perform low-pass filtering on the luminance component image L to obtain the filtered result image Lg. In this embodiment of the invention, the luminance component image L is subjected to Gaussian low-pass filtering to obtain the filtered result image Lg. The purpose is to extract non-detail information from the luminance component image. Using Gaussian low-pass filtering is the best and easiest way to implement.

[0074] The differential transformation module 303 is used to perform differential transformation on the filtered result image Lg using a first preset transformation formula to obtain the difference result image Ld; specifically, the first preset transformation formula is: Ld=Lg-L.

[0075] The nonlinear transformation module 304 is used to perform a nonlinear transformation on the difference result image Ld using a second preset transformation formula to obtain a nonlinear transformed image Lt; specifically, the second preset transformation formula is Lt=L-β*Ld α Where α is the exponential coefficient and β is the linear coefficient, and their values ​​satisfy 0 < α < 1 and 0 < β < 1.

[0076] The enhanced image output module 305 is used to fuse the nonlinear transformed image Lt with other component images, and to perform a second color space conversion operation on the fused image to obtain the output image.

[0077] In one specific embodiment, such as Figure 4 As shown, the image luminance component extraction module 301 includes:

[0078] The normalization unit 3011 is used to normalize the pixel values ​​of the input image, wherein the input image is in RGB color space and the pixel values ​​of the input image after normalization are in the range of [0, 1].

[0079] The first space conversion unit 3012 is used to convert the normalized input image from the RGB color space to the Lab color space.

[0080] The component extraction unit 3013 is used to extract the luminance component image L, the color dimension component image a, and the color dimension component image b corresponding to the Lab color space.

[0081] In one specific embodiment, such as Figure 4 As shown, the enhanced image output module 305 includes:

[0082] The fusion unit 3051 is used to fuse the nonlinear transformed image Lt with the color dimension component image a and the color dimension component image b.

[0083] The second space conversion unit 3052 is used to convert the fused image from Lab color space to RGB color space, and to convert the pixel values ​​of the space-converted image to obtain an output image with pixel values ​​in the range of [0, 255].

[0084] Further functional descriptions of the various modules and units described above are the same as those in the corresponding method embodiments described above, and will not be repeated here. The apparatus in this embodiment is presented in the form of functional units, where a unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0085] This invention also provides a computer device having the above-described features. Figure 3 or Figure 4 The image detail enhancement device shown.

[0086] Please see Figure 5 , Figure 5 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 5As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a set of blade servers, or a multiprocessor system). Figure 5 Take a processor 10 as an example.

[0087] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field-programmable gate array (FPGA), a general-purpose array logic (GPA), or any combination thereof.

[0088] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0089] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device as shown by a landing page for an app. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, which can be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0090] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0091] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0092] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0093] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. An image detail enhancement method, characterized in that, include The input image undergoes a first color space conversion operation, and the luminance component image L and other component images corresponding to the converted color space are extracted. The other component images include color dimension component image a and color dimension component image b. The first color space conversion operation on the input image and the extraction of the luminance component image L and other component images corresponding to the converted color space include: normalizing the pixel values ​​of the input image, wherein the input image is in RGB color space, and the pixel values ​​of the normalized input image are in the range of [0, 1]; converting the normalized input image from the RGB color space to the Lab color space; and extracting the luminance component image L, color dimension component image a, and color dimension component image b corresponding to the Lab color space. Performing a low-pass filter on the luminance component image L to obtain the filtered result image Lg includes: determining the target pixel block corresponding to each pixel in the luminance component image L, wherein the size of the target pixel block is the same as the size of a preset Gaussian low-pass filter kernel template, and the radius of the preset Gaussian low-pass filter kernel template is [missing information]. Let m and n be the length and width of the luminance component image L, respectively, and min(m,n) represent the minimum value of m and n. For each target pixel block, the template is used as a weight and multiplied by the pixel value of each target pixel in the target pixel block, and the resulting values ​​are summed to obtain a replacement pixel value. The replacement pixel value is then used to update the pixel value of the center point of the pixel block. After each pixel in the luminance component image L is updated, the image composed of all the updated pixels is determined as the filtered result image Lg. The filtered image Lg is subjected to a differential transformation using a first preset transformation formula to obtain the difference result image Ld. The first preset transformation formula is: ; The difference result image Ld is nonlinearly transformed using a second preset transformation formula to obtain a nonlinearly transformed image Lt. The second preset transformation formula is: Where α is the exponential coefficient and β is the linear coefficient, and their values ​​satisfy the following conditions: ; The nonlinear transformed image Lt is fused with the other component images, and the fused image is then subjected to a second color space conversion operation to obtain the output image.

2. The image detail enhancement method according to claim 1, characterized in that, The step of fusing the nonlinear transformed image Lt with the other component images, and then performing a second color space conversion operation on the fused image to obtain an output image, includes: The nonlinear transformed image Lt is fused with the color dimension component image a and the color dimension component image b; The fused image is converted from the Lab color space to the RGB color space, and the converted image is then subjected to pixel value conversion to obtain an output image with a pixel value range of [0, 255].

3. An image detail enhancement device, characterized in that, The device includes: An image luminance component extraction module is used to perform a first color space conversion operation on the input image and extract the luminance component image L and other component images corresponding to the converted color space. The other component images include color dimension component image a and color dimension component image b. The image luminance component extraction module includes: The normalization unit is used to normalize the pixel values ​​of the input image, wherein the input image is in RGB color space and the pixel values ​​of the input image after normalization are in the range of [0, 1]. The first space conversion unit is used to convert the normalized input image from the RGB color space to the Lab color space. The component extraction unit is used to extract the luminance component image L, the color dimension component image a, and the color dimension component image b corresponding to the Lab color space. A low-pass filtering module is used to perform low-pass filtering on the luminance component image L to obtain a filtered result image Lg, including: Determine the target pixel block corresponding to each pixel in the luminance component image L, wherein the size of the target pixel block is the same as the size of a preset Gaussian low-pass filter kernel template, and the radius of the preset Gaussian low-pass filter kernel template is 1 / 4. min(m,n), where m and n are the length and width of the luminance component image L, respectively, and min(m,n) represents the minimum value of m and n; for each target pixel block, the template is used as a weight and multiplied by the pixel value of each target pixel in the target pixel block, and the multiplied values ​​are summed to obtain a replacement pixel value, and the pixel value of the center point of the pixel block is updated using the replacement pixel value; after each pixel in the luminance component image L is updated, the image composed of all the updated pixels is determined as the filtered result image Lg; The differential transformation module is used to perform a differential transformation on the filtered result image Lg using a first preset transformation formula to obtain a difference result image Ld. The first preset transformation formula is: ; The nonlinear transformation module is used to perform a nonlinear transformation on the difference result image Ld using a second preset transformation formula to obtain a nonlinear transformed image Lt. The second preset transformation formula is: Where α is the exponential coefficient and β is the linear coefficient, and their values ​​satisfy the following conditions: ; An enhanced image output module is used to fuse the nonlinear transformed image Lt with the other component images, and to perform a second color space conversion operation on the fused image to obtain an output image.

4. A computer device, characterized in that, include: A memory and a processor are communicatively connected, the memory storing computer instructions, and the processor executing the computer instructions to perform the image detail enhancement method according to any one of claims 1 to 2.

5. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing a computer to perform the image detail enhancement method according to any one of claims 1 to 2.

Images