Medical image enhancement method and system based on local information fuzzy reasoning

Abstract

The invention provides a medical image enhancement method and system based on local information fuzzy reasoning, and relates to the technical field of medical images. The method comprises the following steps: firstly, dividing a medical image into a plurality of segmented region sub-blocks, and calculating Michelson contrast and image entropy information of each sub-block; according to the image entropy of each sub-block and the Michelson contrast ratio, determining respective fuzzy sets of the two sub-blocks; then constructing a fuzzy inference system, and taking the image entropy of each sub-block and a fuzzy set of Michelson contrast as the input of the fuzzy inference system to obtain an optimal shear enhancement parameter corresponding to each sub-region; calculating the local shear limitation of each sub-block by using the obtained optimal shear enhancement parameter corresponding to each sub-block; and finally, performing cutting processing on the part, exceeding the cutting limit, of each sub-block, performing histogram equalization, and splicing all the sub-blocks subjected to histogram equalization together to finally obtain a complete enhanced image. According to the method, intelligent and refined contrast enhancement can be realized.

Description

Technical Field

[0001] The present invention relates to the field of medical imaging technology, and in particular, to a medical image enhancement method and system based on local information fuzzy reasoning. Background Art

[0002] X-ray imaging technology is one of the most widely used and effective medical diagnostic methods, and has been used for early detection and prognosis evaluation of various diseases. The medical digital imaging system based on X-rays can clearly present the fat in breast tissue, showing black; masses and calcification deposits are gray; non-cancerous and cancerous masses are shown as white. However, in many cases, due to different internal and external factors, the contrast of the gray levels of these X-ray digital images is often poor and blurred, and this uncertainty in imaging quality will seriously affect the diagnosis of clinicians.

[0003] Contrast refers to the gray level difference between each pixel in an image. As one of the most important features in image processing, good contrast plays a key role in detecting breast masses. The quality of contrast will largely affect the subjective visual quality of the image. However, factors such as the state of the patient during image acquisition, machine resolution, imaging software system, and manual operation may all lead to poor contrast of X-ray digital images. The gray level dynamic range in these images is relatively narrow, so contrast enhancement usually needs to be performed in the post-processing process. During a series of gray level transformation processes, the gray level values will be dynamically stretched within a wider range.

[0004] There are two deficiencies in the existing contrast enhancement methods: the gray level adjustment range of the image is considered from the overall image, and most of the local features of the image are ignored, so the adjustment of image details is often not satisfactory; the predefined limit and the output weight of the defuzzification module for image adjustment need to be set manually, so there is a large degree of randomness. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a medical image enhancement method and system based on local information fuzzy reasoning to enhance the contrast of digital mammography images in view of the above deficiencies of the prior art.

[0006] To solve the above technical problems, the technical solutions adopted by the present invention are as follows:

[0007] On the one hand, the present invention provides a medical image enhancement method based on local information fuzzy reasoning, including:

[0008] Dividing the medical image into several segmented region sub-blocks, and obtaining the Michelson contrast and image entropy information within the range of each sub-block;

[0009] Determine the respective fuzzy sets of the image entropy and Michelson contrast of each sub-block;

[0010] Construct a fuzzy inference system, taking the fuzzy sets of the image entropy and Michelson contrast of each sub-block as the input of the fuzzy inference system to obtain the optimal shear enhancement parameters corresponding to each sub-region;

[0011] Use the optimal shear enhancement parameters corresponding to each sub-block obtained to calculate the local shear limit of each sub-block;

[0012] Perform shearing on the parts of each sub-block that exceed the shear limit and then implement histogram equalization;

[0013] Stitch all the sub-blocks that have undergone histogram equalization together to finally obtain a complete enhanced image.

[0014] Furthermore, the fuzzy inference system includes a fuzzification module and a defuzzification module; the fuzzification module performs inference with the image entropy and Michelson contrast of each sub-block as the input according to the constructed fuzzy rules, and dynamically obtains the fuzzy set of the shear enhancement parameter corresponding to each sub-block; the defuzzification module real-numberizes the fuzzy sets of the shear enhancement parameters corresponding to each sub-block according to the characteristics of the fuzzy sets to obtain the optimal shear enhancement parameters corresponding to the sub-blocks.

[0015] Furthermore, the fuzzy inference system constructs three levels of fuzzy sets of low, medium, and high for the Michelson contrast C, image entropy E, and shear enhancement parameter B respectively, and determines the membership functions of each fuzzy set.

[0016] Furthermore, the formulas of the membership functions of the three fuzzy sets of the Michelson contrast, image entropy, and shear enhancement parameter are as follows:

[0017]

[0018]

[0019] Where: x is the Michelson contrast value of the sub-block, y is the image entropy value of the sub-block, z is the shear enhancement parameter of the sub-block, a is the minimum Michelson contrast, b = (a + c) / 2 is the mean value of the Michelson contrast, c is the maximum Michelson contrast, d is the minimum image entropy, e = (d + f) / 2 is the mean value of the image entropy, f is the maximum image entropy, l is the minimum value of the height of the shear enhancement parameter, m = (l + n) / 2 is the mean value of the height of the shear enhancement parameter, and n is the maximum value of the height of the shear enhancement parameter.

[0020] Further, the method constructs fuzzy rules of the fuzzy inference system by using a triangular membership function. The rules are based on the Michelson contrast x and the image entropy y of the input sub-block, and output fuzzy shear enhancement parameters. An independent inference combined by the Mamdani product operator and the fuzzy union is used to construct a product inference engine, and the centroid defuzzification method is used to obtain the optimal shear enhancement parameter.

[0021] Further, based on the fuzzy union combined inference idea and using the Mamdani product operator, for the fuzzy rule "if the input is E and C, then the output is B", when the input Michelson contrast is x and the image entropy is y, the fuzzy set B of the output shear enhancement parameter is determined. * (z):

[0022]

[0023] Using the centroid defuzzification method, the fuzzy set of the shear enhancement parameter is defuzzified to obtain the optimal shear enhancement parameter α, as shown in the following formula:

[0024]

[0025] where Z is the range of the fuzzy set of the shear enhancement parameter.

[0026] Further, the local shear limit of each sub-block is shown in the following formula:

[0027]

[0028] where m'×n' is the number of pixels in the sub-block area, and L is the number of gray levels of the medical image.

[0029] On the other hand, the present invention also provides a medical image enhancement system based on local information fuzzy inference, including:

[0030] Regional sub-block segmentation module: used to divide the medical image into several segmented regional sub-blocks;

[0031] Sub-block feature extraction module: obtain the Michelson contrast and image entropy information within the range of each sub-block;

[0032] Fuzzy set construction module: determine the respective fuzzy sets of the two according to the image entropy and Michelson contrast of each sub-block;

[0033] Optimal shear enhancement parameter calculation module: construct a fuzzy inference system, use the fuzzy sets of the image entropy and Michelson contrast of each sub-block as the input of the fuzzy inference system, and obtain the optimal shear enhancement parameter corresponding to each sub-region;

[0034] Image enhancement module: Using the optimal shear enhancement parameters corresponding to each sub-block obtained, calculate the local shear limit of each sub-block; perform shearing processing on the parts of each sub-block that exceed the shear limit and then implement histogram equalization, splice all the sub-blocks that have undergone histogram equalization together, and finally obtain a complete enhanced image.

[0035] In a third aspect, the present application proposes an electronic device, including: one or more processors, and a memory, where the memory is used to store instructions, and when the instructions are executed by the one or more processors, the one or more processors are caused to execute the medical image enhancement method based on local information fuzzy inference.

[0036] In a fourth aspect, the present application proposes a computer-readable storage medium, which stores executable instructions, and when the instructions are executed, the processor is caused to execute the medical image enhancement method based on local information fuzzy inference.

[0037] In a fifth aspect, the present application proposes a computer program product, including a computer program or instructions, and when the computer program or instructions are executed by a processor, the medical image enhancement method based on local information fuzzy inference is implemented.

[0038] The beneficial effects produced by adopting the above technical solutions are as follows: A medical image enhancement method and system based on local information fuzzy inference provided by the present invention constructs a new fuzzy inference system, which can automatically and accurately calculate the optimal shear enhancement parameters according to multi-dimensional information such as local area texture, gray scale, and image entropy, realize intelligent and refined contrast enhancement, and provide a new way to improve the diagnostic efficiency of X-ray digital images. Description of the Drawings

[0039] Figure 1 It is a flowchart of a medical image enhancement method based on local information fuzzy inference provided by Embodiment 1 of the present invention;

[0040] Figure 2 It is a comparison diagram of image enhancement results provided by Embodiment 1 of the present invention. Detailed Description of the Embodiments

[0041] The following will further describe in detail the specific embodiments of the present invention in conjunction with the drawings and embodiments. The following embodiments are used to illustrate the present invention, but are not used to limit the scope of the present invention.

[0042] Embodiment 1:

[0043] In this embodiment, a medical image enhancement method based on local information fuzzy inference, as Figure 1 shown, includes the following steps:

[0044] Step 1: Obtain a mammogram as the source image and perform preprocessing operations on the source image;

[0045] Step 2: Image segmentation: Segment the low-contrast mammogram into M×N sub-block regions of equal size (size S×S);

[0046] Step 3: Feature calculation: Extract the local feature information of each sub-block, and calculate the contrast and image entropy of each sub-block;

[0047] The main disadvantage of the existing limited contrast histogram equalization extension method is that the histogram clipping enhancement parameter will be automatically set to an unstable value and the characteristics of each window region are ignored. Inappropriate clipping enhancement parameters will seriously affect the contrast enhancement effect, and the predefined clipping enhancement parameters cannot be generalized to different types of data sets. Therefore, in the present invention, through a fuzzy inference system, the clipping enhancement parameter B of each sub-block is dynamically adjusted based on the fuzzy combination of the Michelson contrast C and the image entropy E;

[0048] The Michelson contrast of each sub-block is expressed as:

[0049]

[0050] where max 1≤i≤I x i and min 1≤i≤I x i are respectively the maximum and minimum values of the pixel gray levels in the sub-block of the segmented region, and I is the number of pixels in the sub-block;

[0051] And the image entropy of each sub-block is expressed as:

[0052]

[0053] where p(x i ) is the probability that the i-th pixel gray level x i appears in the sub-block of the segmented region;

[0054] Step 4: Fuzzification processing: Construct a fuzzy inference system, input the contrast and image entropy of the current sub-block into the constructed fuzzy inference system for fuzzification processing, and determine the fuzzy set of the clipping enhancement parameter of the current sub-block according to the established inference rules;

[0055] The fuzzy inference system includes a fuzzification module and a defuzzification module; the fuzzification module performs inference with the image entropy and Michelson contrast of each sub-block as inputs according to the constructed fuzzy rules, and dynamically obtains the fuzzy set of the shear enhancement parameter for each sub-block; the defuzzification module real-numberizes the fuzzy set of the shear enhancement parameter corresponding to each sub-block according to the characteristics of the fuzzy set to obtain the optimal shear enhancement parameter corresponding to the sub-block; its advantage is that when the image pixels are concentrated in a relatively narrow histogram region, the image entropy is relatively low; while when the pixels are evenly distributed over the entire histogram, the image entropy is relatively high.

[0056] In this embodiment, low, medium, and high-level fuzzy sets are respectively constructed for the Michelson contrast C, image entropy E, and shear enhancement parameter B, and the membership functions corresponding to the membership ranges of each fuzzy set are determined.

[0057] Table 1 shows the three fuzzy sets of the Michelson contrast C, which are respectively represented by low (C L ), medium (C M ), and high (C H ), and their membership ranges. Table 2 shows the three fuzzy sets of the image entropy E, which are respectively represented by low (E L ), medium (E M ), and high (E H ), and their membership ranges. Table 3 shows the three fuzzy sets of the shear enhancement parameter B, which are respectively represented by low (B L ), medium (B M ), and high (B H ), and their membership ranges.

[0058] Table 1 Membership Ranges of Michelson Contrast

[0059]

[0060]

[0061] Table 2 Membership Ranges of Image Entropy

[0062]

[0063] Table 3 Membership Ranges of Shear Enhancement Parameter

[0064]

[0065] In this embodiment, the membership function formulas of the three fuzzy sets of the Michelson contrast, image entropy, and shear enhancement parameter are respectively as follows:

[0066]

[0067]

[0068] Among them: x is the Michelson contrast value of the sub-block, y is the image entropy value of the sub-block, z is the shear enhancement parameter of the sub-block, a is the minimum Michelson contrast, b = (a + c) / 2 is the average Michelson contrast, c is the maximum Michelson contrast, d is the minimum image entropy, e = (d + f) / 2 is the average image entropy, f is the maximum image entropy, l is the minimum value of the shear enhancement parameter height, m = (l + n) / 2 is the average shear enhancement parameter height, and n is the maximum value of the shear enhancement parameter height.

[0069] Step 5: Defuzzification to obtain the parameter: After obtaining the fuzzy set of the shear enhancement parameter of the current sub-block, perform the defuzzification operation on the fuzzy set of the shear enhancement parameter, so as to obtain the optimal shear enhancement parameter α corresponding to the current sub-block;

[0070] In this embodiment, a triangular membership function is used to construct the fuzzy rules of the fuzzy inference system. This rule is based on the Michelson contrast x and the image entropy y of the input sub-block, and outputs a fuzzy shear enhancement parameter. An independent inference combined by the Mamdani product operator and the fuzzy union combination is used to construct a product inference engine, and the centroid defuzzification method is used to obtain the optimal shear enhancement parameter.

[0071] First, determine 9 fuzzy inference rules according to Table 4:

[0072] Table 4 Fuzzy Inference Rules

[0073]

[0074] Secondly, based on the fuzzy union combination inference idea, using the Mamdani product operator, for the fuzzy rule "if the input is E and C, then the output is B", when the input entropy is y and the contrast is x, determine the fuzzy set B of the output shear enhancement parameter * (z):

[0075]

[0076] Finally, use the centroid defuzzification method to defuzzify the fuzzy set of the shear enhancement parameter to obtain the optimal shear enhancement parameter α, as shown in the following formula:

[0077]

[0078] Among them, Z is the range of the fuzzy set of the shear enhancement parameter;

[0079] Table 4 above shows nine rules generated based on the Michelson contrast and image entropy of the input data, which are used to determine the most appropriate shear enhancement parameter grading level for each window image. Then, through defuzzification, it is converted into the required shear enhancement parameter.

[0080] Step 6: Sub-block processing: Use the optimal shear enhancement parameter α corresponding to the obtained sub-block to calculate the local shear limit of the sub-block; subsequently, perform histogram equalization on the sub-block and clip the part exceeding the shear limit, which can effectively avoid the problem of over-enhancement according to the local characteristics of the current sub-block and retain image details.

[0081] The local shear limit of each sub-block is shown in the following formula:

[0082]

[0083] where m'×n' is the number of pixels in the sub-block area and L is the gray level number of the medical image.

[0084] Step 7: Loop processing: Repeat Steps 3 to 6 until all sub-blocks have completed the equalization process.

[0085] Step 8: Image stitching: Stitch all the sub-blocks that have undergone histogram equalization together to obtain the complete enhanced image.

[0086] The medical image enhancement method of the present invention can adaptively select the most appropriate shear enhancement parameter for each sub-block in the low-contrast mammogram within a local range, and then perform contrast enhancement. This process effectively avoids the problems of over-enhancement and detail loss, and significantly improves the visual quality of the image. To evaluate the quality of the enhanced image, this embodiment uses a variety of image quality assessment tools to comprehensively quantify and evaluate the enhanced image. Specifically, this embodiment uses indicators such as image entropy, Natural Image Quality Evaluator (NIQE), and Patch-based Contrast Quality Index (PCQI) to carefully analyze and evaluate the enhanced image. These indicators not only cover the richness of image details and information volume, but also comprehensively consider the overall quality of the image from the perspectives of natural visual quality and local contrast.

[0087] To more intuitively demonstrate the superiority of the method of the present invention, this embodiment compares and analyzes the evaluation results of these indicators with classical image enhancement methods. By comparing the results of enhancing ten groups of low-contrast mammograms by different methods, it can be clearly seen that the method of the present invention has significant advantages over traditional methods in improving image contrast, detail performance, and overall visual quality. The specific experimental result comparison is as Figure 2 shown.

[0088] To comprehensively evaluate the quality of the enhanced image, this embodiment also uses quality measurement tools such as image entropy, NIQE, and PCQI to evaluate the enhanced image. After the measurement is completed, the values obtained by these tools are compared and analyzed with the results of classical image enhancement methods, as shown in Tables 5, 6, and 7.

[0089] Among many existing contrast enhancement methods, Histogram Equalization (HE) is one of the most common. Its main idea is to enhance the contrast of an image by roughly evenly redistributing the histogram of the image. However, HE directly processes the global image without considering local information. Adaptive Histogram Equalization (AHE) is based on HE and pays more attention to local contrast. It divides the image into blocks and performs histogram equalization on each sub-block, but this also leads to over-enhancement and some noises will be amplified. Brightness Preserving Dynamic Histogram Equalization (BPDHE) is an extended method of dynamic histogram equalization and brightness preserving multi-modal histogram equalization. This technique divides the input histogram based on local maxima, but it is not suitable for enhancing low-brightness images. Contrast Limited Adaptive Histogram Equalization (CLAHE) is a method that enhances the contrast of an image while reducing noise by clipping the image histogram with a predefined clipping enhancement parameter. Since the clipping enhancement parameter needs to be set in advance and is usually regarded as a constant, classical CLAHE does not perform well for different images. To solve this problem, S. Vidya et al. proposed a CLAHE algorithm based on triangular fuzzy membership functions, which replaces the explicit clipping with fuzzy clipping and preserves the brightness and naturalness of the image. Metrics such as NIQE are often used to prove the effectiveness of the enhanced image. S. Jenifer et al. proposed a CLAHE algorithm based on fuzzy logic. This method constructs an inference system, divides the clipping into three levels: low, medium, and high, uses the contrast and entropy of the image as the basis for classification, and then compares with various existing methods according to image quality evaluation criteria such as entropy to prove the effectiveness of the method. R. Jebadass and P. Balasubramaniam proposed a new method that combines contrast limited adaptive histogram equalization with interval-valued intuitionistic fuzzy sets to enhance color images taken under low light. The interval-valued intuitionistic fuzzy set based on the intuitionistic fuzzy set constructed by fuzzy sets is used to enhance images taken under low light. In this method, first, the given low-light image is fuzzified by conventional fuzzification; then the fuzzified image is converted into an interval-valued intuitionistic fuzzy image; finally, based on image quality measurement tools such as entropy and correlation, it is compared with various existing methods, and the results show that this method has better effects. P. Mamoria and D. Raj proposed a new method based on the Mamdani fuzzy inference system and using multiple fuzzy membership functions. First, the input image is transformed into the fuzzy domain. Then, a set of fuzzy "if-then" rule bases in the fuzzy domain gives the best results in image contrast enhancement. Based on different combinations of the shapes of the membership functions, an optimal prediction solution can be determined, which can be applied to different types of input images according to application requirements. The result analysis shows that in the fuzzy domain, quality attributes such as PCQI will show different performance with the selection of the number and size of the membership functions.This study proposed an algorithm for determining the optimal combination of fuzzy membership functions to achieve optimal image contrast enhancement. V. Magudeeswaran et al. proposed contrast-limited fuzzy adaptive histogram equalization for enhancing the contrast of brain magnetic resonance imaging (MRI) images. This method is divided into three stages: First, the gray intensity is transformed into the membership plane, and the membership plane is modified using a contrast enhancement operator; Second, the CLAHE method is applied to the modified membership plane to enhance the image contrast and retain the original brightness; Finally, the membership plane is mapped back to the gray intensity, and it is compared and evaluated with existing methods according to image quality metrics such as entropy. The results show that this method significantly enhances the contrast of digital images and better retains the original brightness. B. S. Min et al. proposed a method for determining CLAHE parameters based on image entropy. By fitting the image entropy curve, the point with the maximum curvature curve value is found as the clipping limit. Then, the experimental results are analyzed using the average intensity, root mean square contrast, and entropy. The results show better subjective performance. L. Chandrashekar et al. proposed a multi-objective enhancement technique with entropy, edge intensity, and the sum of edge pixels as the objective function. The particle swarm optimization algorithm is used to find the operation clipping limit that can enhance MRI images. This technique shows better enhancement effects in terms of contrast, entropy, etc. J. Joseph et al. proposed to select the optimal clipping enhancement parameters based on the maximum value of image entropy and the minimum value of SSIM. Statistical metrics such as PCQI that can quantify the degree of feature degradation and contrast enhancement are used to objectively determine the optimal parameter settings of CLAHE for MRI images.

[0090] Image entropy is mainly used to measure the richness of information contained in an image after enhancement. Although image entropy does not directly represent an improvement in image quality, it can depict the richness of details.

[0091] Table 5 Image Entropy Evaluation Results

[0092]

[0093] PCQI is an important metric commonly used in image quality assessment, mainly focusing on evaluating the contrast quality of images. Different from traditional image quality assessment methods, PCQI fully considers the contrast and brightness differences in each local region of the image. By integrating these factors, a score that comprehensively reflects the image quality is calculated. In the specific calculation process, PCQI uses non-overlapping image blocks for operation. For each image block, the contrast within the block and the brightness difference between blocks are calculated respectively. Based on these calculation results, the overall contrast quality of the image is further obtained. Due to its comprehensiveness and accuracy in evaluation, PCQI has been widely used in fields such as image processing and compression, and is often used to compare the actual effects of different image processing algorithms and compression coding schemes. Generally, the larger the PCQI index, the better the enhancement effect of the image. The definition formula of PCQI is as follows:

[0094]

[0095] where N is the number of blocks, w j is the weight of the j-th image block, μ j is the mean value of the j-th image block, is the mean value of the j-th reference image block, σ j is the variance of the j-th image block, is the variance of the j-th reference image block, and c1, c2 are constants.

[0096] Table 6 PCQI evaluation results

[0097]

[0098] NIQE is a metric for evaluating image quality, with the full name of No-Reference Image Quality Evaluator. It evaluates image quality based on the statistical regularity and texture complexity in natural images, and can be used to evaluate the quality of distorted compressed images as well as other types of image distortions. The larger the value of the NIQE metric, the worse the image quality.

[0099]

[0100] where v1, v2, ∑1, ∑2 represent the mean vector and covariance matrix respectively.

[0101] Table 7 NIQE evaluation results

[0102]

[0103]

[0104] By comparing the enhancement results, it can be seen from the experimental results that the enhancement effect obtained by using the method of the present invention is not only better but also more stable than other histogram equalization methods. The medical images used in this embodiment contain many fine lines. Therefore, the algorithm combines the local feature information of the sub-blocks for processing, and thus obtains a more ideal enhancement result. It can be clearly seen from the experimental data table that when using the method of the present invention for enhancement, both the stability of the scores of the image enhancement evaluation indicators and the actual enhancement effect are better than other contrast enhancement algorithms of histogram equalization.

[0105] Embodiment 2:

[0106] This embodiment provides a medical image enhancement system based on local information fuzzy inference, including:

[0107] Region sub-block segmentation module: used to divide the medical image into several segmented region sub-blocks;

[0108] Sub-block feature extraction module: obtain the Michelson contrast and image entropy information within the range of each sub-block;

[0109] Fuzzy set construction module: determine the respective fuzzy sets of the two according to the image entropy and Michelson contrast of each sub-block;

[0110] Optimal shear enhancement parameter calculation module: construct a fuzzy inference system, use the fuzzy sets of the image entropy and Michelson contrast of each sub-block as the input of the fuzzy inference system, and obtain the optimal shear enhancement parameter corresponding to each sub-region;

[0111] Image enhancement module: use the obtained optimal shear enhancement parameter corresponding to each sub-block to calculate the local shear limit of each sub-block; perform shear processing on the part of each sub-block that exceeds the shear limit and then implement histogram equalization, splice all the sub-blocks after histogram equalization together, and finally obtain a complete enhanced image.

[0112] Embodiment 3:

[0113] This embodiment proposes an electronic device, including: one or more processors, and a memory, where the memory is used to store instructions, and when the instructions are executed by the one or more processors, the one or more processors execute the medical image enhancement method based on local information fuzzy inference.

[0114] The electronic device can be a mobile phone, a computer, a tablet computer, etc., including a memory and a processor. A computer program is stored on the memory, and when the computer program is executed by the processor, it implements the medical image enhancement method based on local information fuzzy inference as described in the embodiments. It can be understood that the electronic device may further include an input / output (I / O) interface and a communication component.

[0115] Among them, the processor is used to execute all or part of the steps in the medical image enhancement method based on local information fuzzy inference as described in the above embodiments. The memory is used to store various types of data, which may include, for example, instructions of any application program or method in the electronic device, and data related to the application program.

[0116] The processor can be implemented by an application specific integrated circuit (ASIC), a digital signal processor (DSP), a programmable logic device (PLD), a field programmable gate array (FPGA), a controller, a microcontroller, a microprocessor or other electronic components, and is used to execute the medical image enhancement method based on local information fuzzy inference as described in the above embodiments.

[0117] Embodiment 4:

[0118] This embodiment provides a computer-readable storage medium, which stores executable instructions. When the instructions are executed and implemented in the form of a software functional unit and sold or used as an independent product, they can be stored in a computer-readable storage medium.

[0119] The computer software product is stored in a storage medium, including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the medical image enhancement method based on local information fuzzy inference described in various embodiments of the present application.

[0120] The aforementioned storage medium includes: flash memory, hard disk, multimedia card, card-type memory (such as SD (Secure Digital Memory Card), or DX (abbreviation for Memory Data Register, MDR), memory data register, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, server, APP (abbreviation for Application, application software) application mall, and various other media that can store program verification codes. A computer program is stored thereon, and when the computer program is executed by a processor, each step of the above-mentioned medical image enhancement method based on local information fuzzy inference can be implemented.

[0121] Embodiment 5:

[0122] This embodiment provides a computer program product, including a computer program or instruction, which implements the above-mentioned medical image enhancement method based on local information fuzzy inference when executed by a processor.

[0123] Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a computer program product.

[0124] Each embodiment in this application is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments.

[0125] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope defined by the claims of the present invention.

Claims

1. A medical image enhancement method based on local information fuzzy inference, characterized in that: Comprising: Dividing a medical image into a number of segmented region sub - blocks, and obtaining the Michelson contrast and image entropy information within the range of each sub - block; Determining the respective fuzzy sets of the image entropy and Michelson contrast for each sub - block; Constructing a fuzzy inference system, taking the fuzzy sets of the image entropy and Michelson contrast of each sub - block as the input of the fuzzy inference system, and obtaining the optimal shear enhancement parameter corresponding to each sub - region; Calculating the local shear limit of each sub - block by using the obtained optimal shear enhancement parameter corresponding to each sub - block; Performing a shearing process on the part of each sub - block that exceeds the shear limit, and implementing histogram equalization; Stitching all the sub - blocks after histogram equalization together to finally obtain a complete enhanced image.

2. The medical image enhancement method based on local information fuzzy inference according to claim 1, characterized in that: The fuzzy inference system includes a fuzzification module and a defuzzification module; the fuzzification module, based on the constructed fuzzy rules, takes the image entropy and Michelson contrast of each sub - block as the input for inference, and dynamically obtains the fuzzy set of the shear enhancement parameter corresponding to each sub - block; the defuzzification module real - numberizes the fuzzy set of the shear enhancement parameter corresponding to each sub - block according to the characteristics of the fuzzy set to obtain the optimal shear enhancement parameter corresponding to the sub - block.

3. The medical image enhancement method based on local information fuzzy inference according to claim 2, characterized in that: The fuzzy inference system constructs three levels of fuzzy sets, namely low, medium, and high, for the Michelson contrast C, image entropy E, and shear enhancement parameter B respectively, and determines the membership functions of each fuzzy set.

4. The medical image enhancement method based on local information fuzzy inference according to claim 3, characterized in that: The membership function formulas of the three fuzzy sets of the Michelson contrast, image entropy, and shear enhancement parameter are as follows: Where: x is the Michelson contrast value of the sub - block, y is the image entropy value of the sub - block, z is the shear enhancement parameter of the sub - block, a is the minimum Michelson contrast, b=(a + c) / 2 is the mean value of the Michelson contrast, c is the maximum Michelson contrast, d is the minimum image entropy, e=(d + f) / 2 is the mean value of the image entropy, f is the maximum image entropy, l is the minimum value of the height of the shear enhancement parameter, m=(l + n) / 2 is the mean value of the height of the shear enhancement parameter, and n is the maximum value of the height of the shear enhancement parameter.

5. The medical image enhancement method based on local information fuzzy inference according to claim 4, characterized in that: The method constructs the fuzzy rules of the fuzzy inference system by using a triangular membership function. This rule, based on the Michelson contrast x and image entropy y of the input sub - block, outputs a fuzzy shear enhancement parameter, constructs a product inference engine by using independent inference combined with the Mamdani product operator and fuzzy union combination, and uses the centroid defuzzification method to obtain the optimal shear enhancement parameter; Based on the idea of fuzzy union combination reasoning and using the Mamdani product operator, for the fuzzy rule "if the input is E and C, then the output is B", when the input Michelson contrast is x and the image entropy is y, determine the fuzzy set B of the output shear enhancement parameter * (z): Using the centroid defuzzification method to defuzzify the fuzzy set of the shear enhancement parameter to obtain the optimal shear enhancement parameter α, as shown in the following formula: Where, Z is the range of the fuzzy set of the shear enhancement parameter.

6. The medical image enhancement method based on local information fuzzy inference according to claim 5, characterized in that: The local shear limit of each sub - block is as shown in the following formula: Where, m'×n' is the number of pixels in the sub - block area, and L is the number of gray levels of the medical image.

7. A medical image enhancement system based on local information fuzzy inference, implemented based on the method described in claim 1, characterized in that: Comprising: A regional sub - block segmentation module: used for dividing a medical image into a number of segmented region sub - blocks; A feature extraction module: obtaining the Michelson contrast and image entropy information within the range of each sub - block; Fuzzy set construction module: Determine the respective fuzzy sets of the image entropy and Michelson contrast of each sub-block; Optimal shear enhancement parameter calculation module: Construct a fuzzy inference system, use the fuzzy sets of the image entropy and Michelson contrast of each sub-block as the input of the fuzzy inference system, and obtain the optimal shear enhancement parameter corresponding to each sub-region; Image enhancement module: Use the obtained optimal shear enhancement parameter corresponding to each sub-block to calculate the local shear limit of each sub-block; perform histogram equalization after shearing the part of each sub-block that exceeds the shear limit, and splice all the sub-blocks that have undergone histogram equalization together to finally obtain a complete enhanced image.

8. An electronic device, comprising: One or more processors, and a memory, the memory is used to store instructions, when the instructions are executed by the one or more processors, the one or more processors execute the medical image enhancement method based on local information fuzzy inference described in claim 1.

9. A computer-readable storage medium, which stores executable instructions, and when the instructions are executed, the processor executes the medical image enhancement method based on local information fuzzy inference described in claim 1.

10. A computer program product, including a computer program or instructions, which when executed by a processor implements the medical image enhancement method based on local information fuzzy inference described in claim 1.

Images