A method for laparoscopic image enhancement optimization

Abstract

The invention discloses a method for laparoscopic image enhancement optimization, which belongs to the field of image processing. The steps include: S1, in the initialization stage of the mountain gazelle optimization algorithm, the algorithm population is initialized by a good point set strategy, and the dolphin position update strategy in the initialization stage of the algorithm is improved; S2, the search performance control factor a of the mountain gazelle optimization algorithm is improved by a segmented method, and the algorithm search performance parameter Cof is improved. i S3. Introducing the "variable convergence" mechanism in the convergence process of the Mountain Gazelle Optimization Algorithm to improve the optimization mechanism of the Mountain Gazelle Optimization Algorithm; S4. Using the Improved Mountain Gazelle Optimization Algorithm (IMGO) to adjust the correction coefficient α of the MSRCR algorithm, and establishing the IMGO‑MSRCR algorithm model; S5. Inputting the salient point set of the abdominal image into the IMGO‑MSRCR algorithm model, and performing contrast enhancement processing on the abdominal image according to the information of the salient points.

Description

Technical Field

[0001] The invention relates to the field of image processing, and in particular to a method for optimizing laparoscopic image enhancement. Background Art

[0002] High-definition laparoscopic images help doctors better observe the pathological tissue, anatomical structure and vascular distribution, so as to perform surgical operations more accurately and reduce the risk of surgical complications and misoperation. By observing laparoscopic images, doctors can preliminarily judge the nature, degree and scope of the lesion, which helps doctors make more accurate diagnoses and treatment plans during surgery.

[0003] The MSRCR algorithm is based on the Retinex theory, using multi-scale analysis and the principle of color constancy to enhance the image through a series of processing processes, so that it can achieve a balance in edge enhancement, dynamic range compression and color constancy. This algorithm can adaptively process various types of images and improve the visual effect of the image; the MSRCR algorithm decomposes the image into sub-images of different scales through multi-scale analysis, and then enhances the sub-images using the principle of color constancy, and finally fuses the enhanced sub-images to obtain the final output image.

[0004] The color restoration factor parameter c of the MSRCR algorithm is used to adjust the proportional relationship between the three color channels in the original image to eliminate the defect of image color distortion. If the parameter c value is set improperly, color deviation or distortion will occur. Therefore, it is necessary to select a suitable parameter c value according to the specific image content and application scenario to obtain a better color restoration effect. The setting of the correction coefficient α value can affect the contrast and brightness of the image, thereby affecting the overall visual effect of the image. The color restoration factor parameter c and correction coefficient α in the traditional MSRCR algorithm are difficult to adapt automatically and need to be adjusted manually, resulting in color deviation or distortion in the enhanced image.

[0005] The Mountain Gazelle Optimizer (MGO) algorithm is a meta-heuristic optimization algorithm that establishes a mathematical model by simulating the social habits of mountain gazelles for optimization problems. The Mountain Gazelle Optimizer (MGO) algorithm has good search capabilities and shows good performance even when the dimension of the optimization problem is increased. However, the global search and local development balance performance of the Mountain Gazelle optimization algorithm are poor, and the algorithm's optimization accuracy needs to be improved. Summary of the invention

[0006] In response to the above-mentioned technical problems, the present technical scheme provides a method for laparoscopic image enhancement optimization. By improving the Mountain Gazelle Optimization Algorithm (MGO), the balance between the global search and local development of the MGO algorithm is improved, thereby improving the optimization speed of the MGO algorithm while avoiding the MGO algorithm from falling into the local optimum. The correction coefficient α of the MSRCR algorithm is adjusted by using the improved Mountain Gazelle Optimization Algorithm (IMGO), and the IMGO-MSRCR algorithm is obtained to process abdominal images, thereby improving the detail display effect of the abdominal images.

[0007] To achieve the above object, the technical solution of the present invention is:

[0008] A method for laparoscopic image enhancement optimization, the specific steps comprising:

[0009] S1. In the initialization stage of the mountain gazelle optimization algorithm, the algorithm population is initialized through the good point set strategy, and the dolphin position update strategy in the initialization stage of the algorithm is improved.

[0010] S2. Improve the search performance control factor a of the Mountain Gazelle optimization algorithm by a segmented method, and improve the algorithm search performance parameter Cof i The mathematical model of .

[0011] S3. Introduce the "variable convergence" mechanism in the convergence process of the mountain gazelle optimization algorithm to improve the optimization mechanism of the mountain gazelle optimization algorithm.

[0012] S4. Use the improved mountain gazelle optimization algorithm (IMGO) to adjust the correction coefficient α of the MSRCR algorithm and establish the IMGO-MSRCR algorithm model.

[0013] S5. Input the salient point set of the abdominal cavity image into the IMGO-MSRCR algorithm model, and perform contrast enhancement processing on the abdominal cavity image according to the information of the salient points.

[0014] Preferably, the position update strategy of the original initialization stage of the mountain gazelle optimization algorithm is implemented by a random initialization method. Random initialization will cause the population individuals of the mountain gazelle optimization algorithm to be very concentrated in some areas and very dispersed in some areas, which will increase the probability of the algorithm falling into a local optimal solution too early in the search process, and reduce the stability of the mountain gazelle optimization algorithm; and the initialization of the good point set is more evenly distributed than random nodes, which makes the mountain gazelle optimization algorithm have better traversal, which is conducive to the algorithm to obtain the best global optimization effect and improve the optimization accuracy of the mountain gazelle optimization algorithm.

[0015] Furthermore, in the initial stage of the mountain gazelle optimization algorithm, the initialized mountain gazelle represents the candidate solution of the problem to be optimized, and the position of the mountain gazelle represents the problem variable, that is, the correction coefficient α solution set of the MSRCR algorithm. The correction coefficient α solution set of the MSRCR algorithm is encoded into the individual position of the mountain gazelle updated in the optimization process of the mountain gazelle optimization algorithm through the real number coding method. The mathematical model formula is:

[0016]

[0017] In the formula, is the initial position value of the j-th dimension of the i-th mountain gazelle, lb j is the lower limit of the position of the j-th dimension, ub j is the upper limit of the position of the j-th dimension, P N (i) is a set of good points with the number N, where N is the population size of the mountain gazelle optimization algorithm.

[0018] Preferably, the search performance control factor a of the Mountain Gazelle optimization algorithm affects the search performance parameter Cor i , which determines the optimization speed of the mountain gazelle optimization algorithm. The search performance control factor a before the improvement adopts a linear change method to change with the number of iterations. Regardless of the early or late algorithm iterations, the same global search speed and local development speed are maintained, resulting in the global search and local development performance of the mountain gazelle optimization algorithm cannot be well balanced; the search performance control factor a is improved in a piecewise manner, so that its value in the early stage remains large, thereby ensuring that the algorithm is mainly based on global search in the early stage, and the value in the later stage remains small, thereby ensuring that the algorithm is mainly based on convergence in the later stage, so that the global search and local development are better balanced.

[0019] Furthermore, the improved search performance control factor a 2 The mathematical model formula is:

[0020]

[0021] In the formula, a max is the maximum value of the search performance control factor, which is 1.2. min To search for the minimum value of the performance control factor, the value is 0.2.

[0022] Preferably, a "variable convergence" mechanism is introduced in the convergence process of the mountain gazelle optimization algorithm. After the position update of the mountain gazelle optimization algorithm is completed, when the mountain gazelle optimization algorithm falls into a local optimum, the local optimum is jumped out through the "variable convergence" mechanism to improve the optimization accuracy of the mountain gazelle optimization algorithm. Specifically, the variable neighborhood search algorithm is integrated. When a solution better than the current solution is not found in the current neighborhood search, the mountain gazelle optimization algorithm jumps to the next neighborhood to continue searching; when a solution better than the current solution is found in the current neighborhood search, it jumps back to the first neighborhood and starts searching again.

[0023] Preferably, the mathematical model formula of the MSRCR algorithm is:

[0024]

[0025] In the formula, R MSRCR is the enhanced abdominal image, c i (x, y) is the color restoration factor of the i-th pixel of the original abdominal image, R is the total number of pixels, α is the correction coefficient of the MSRCR algorithm, S i (x, y) is the original pixel image of the abdominal cavity, F i (x, y) is the abdominal cavity pixel image after filtering.

[0026] Preferably, the original abdominal cavity image is preprocessed by low-pass filtering using a Gaussian function to obtain a filtered abdominal cavity pixel image F i (x, y).

[0027] Preferably, a significant point set is extracted from the preprocessed abdominal image, and the color restoration factor c is adjusted according to the information of the significant point set; for a region with high significance, the c value is appropriately increased to enhance the contrast of the region, and for a region with low significance, the c value is appropriately reduced to maintain the color stability of the region; the original abdominal image S i (x, y) and the filtered abdominal image F i (x, y) is converted to the logarithmic domain and the difference is performed to obtain the reflection image in the logarithmic domain. The reflection images of multiple scales are weighted summed, and finally, the enhanced abdominal image is output.

[0028] Preferably, when the MSRCR processes an image, there is color distortion. When the correction coefficient α of the MSRCR algorithm is adjusted by the improved mountain gazelle optimization algorithm (IMGO), the performance of the correction coefficient α of the MSRCR algorithm is reflected according to the contrast and Tenengrad. The fitness value function of the improved mountain gazelle optimization algorithm (IMGO) when adjusting the correction coefficient α of the MSRCR algorithm is designed by taking the sum of the contrast value and the reciprocal of the Tenengrad value as the objective function J. The fitness function is designed by the minimum value of the objective function. The smaller the fitness function value is, the more the correction coefficient α of the MSRCR algorithm is adjusted by the improved mountain gazelle optimization algorithm (IMGO). The requirements of MSRCR in processing abdominal images are more met; the mathematical model formula of the objective function J is:

[0029]

[0030] Where max(σ) is the maximum standard deviation of the abdominal image, μ is the mean of the abdominal image, MN is the total number of pixels in the abdominal image, M is the number of horizontal pixels in the abdominal image, N is the number of vertical pixels in the abdominal image, and x is m (u, v) and x n (u, v)) is the difference between the pixels in the horizontal and vertical directions of the abdominal image pixel (u, v).

[0031] Preferably, the specific steps of adjusting the correction coefficient α of the MSRCR algorithm using the improved mountain gazelle optimization algorithm (IMGO) are:

[0032] S41. Improve the mountain gazelle optimization algorithm through good point set strategy initialization, including the population size NP of the mountain gazelle algorithm, the problem dimension D, and the initial position of the mountain gazelle The lower limit of the position of the jth dimension of the algorithm lb j and the j-th dimension position upper limit ub j And the maximum number of iterations T, the maximum value of the search performance control factor a max and search performance control factor minimum value a min ;

[0033] S42, encoding the correction coefficient α of the mathematical model of the MSRCR enhanced abdominal image as the location of the individual mountain gazelle;

[0034] S43, simulate the social behavior of mountain gazelles, establish the algorithm mathematical model in four stages, update the individual positions of mountain gazelle population, and the update of individual positions of mountain gazelles is the update of the correction coefficient α solution of MSRCR;

[0035] S44, calculating the fitness value of the individual position of the mountain gazelle population after this iteration update through the objective function, and retaining the individual position of the mountain gazelle population corresponding to the minimum fitness value;

[0036] S45, compare the minimum fitness value of the current iteration with the fitness value of the previous iteration, calculate the speed reduction ratio q of the fitness values ​​before and after, when q≤0.27, it means that the algorithm optimization falls into the global suboptimal solution, use the "change convergence" mechanism to change the neighborhood range, and update the population position to jump out of the global suboptimal solution;

[0037] S46. Use the greedy selection strategy to retain the individual position of the mountain gazelle population with the smallest global fitness value. The current iteration t=t+1. Determine whether the current iteration t is greater than the total number of iterations T. If so, end the algorithm optimization and output the global optimal individual position of the mountain gazelle population. After analysis, assign it to the correction coefficient α of the mathematical model of MSRCR enhanced abdominal images.

[0038] Furthermore, in said S43, the social behavior of mountain gazelles is simulated, an algorithm mathematical model is established in four stages, and the individual positions of the mountain gazelle population are updated, and the specific steps are as follows:

[0039] S431, the improved algorithm searches for performance parameters Cof i After segmentation, the mathematical model was applied to the four stages of mountain gazelle social behavior, and the improved algorithm search performance parameter Cof i The mathematical model formula is:

[0040]

[0041] In the formula, r 3 and r 4 is a random number between 0 and 1, N 2 (D), N 3 (D) and N 4 (D) is a random number within the dimension D of the algorithm problem;

[0042] S442, executing the first phase position update strategy of the improved mountain gazelle optimization algorithm (IMGO), and updating the individual positions of the mountain gazelle population according to formula (1):

[0043] TSM=X best -|(r 1 ×BH-r 2 ×X(t))×F|×Cof 1 (1);

[0044] Where TSM is the position of the first stage of the Mountain Gazelle optimization algorithm, X best is the global optimal position of the current iteration, r 1 and r 2is a random number between 0 and 1, BH is the individual coefficient that has the ability to reach the optimal position, X(t) is the position of the mountain gazelle in the current iteration, F is the weight coefficient of the position update strategy in the first stage of the mountain gazelle optimization algorithm, Cof 1 Searching for performance parameters for the improved algorithm of the first phase;

[0045] S443, executing the position update strategy of the second phase of the improved mountain gazelle optimization algorithm (IMGO), and updating the individual positions of the mountain gazelle population according to formula (2):

[0046] MH=(BH+Cof 2 )+(r 3 ×X best -r 4 ×X rand )×Cof 2 ;

[0047] Where MH is the position of the second stage of the Mountain Gazelle optimization algorithm, Cof 2 Search performance parameters for the improved algorithm in the second phase, X rand is the vector position of a gazelle randomly selected from the entire population;

[0048] S444, executing the position update strategy of the third phase of the improved mountain gazelle optimization algorithm (IMGO), and updating the individual positions of the mountain gazelle population according to formula (3):

[0049] BMH=(X(t)-W)+(r 5 ×X best -r 6 ×BH)×Cof 3 ;

[0050] Where BMH is the position of the mountain gazelle in the third stage of the optimization algorithm, W is the control factor of the mountain gazelle on the territory, and r 5 and r 6 is a random number between 0 and 1, Cof 3 Searching for performance parameters for the improved algorithm in the third phase;

[0051] S445, executing the fourth stage position update strategy of the improved mountain gazelle optimization algorithm (IMGO), and updating the individual positions of the mountain gazelle population according to formula (4):

[0052] MSF=(ub j -lb j )×Cof 4 +lb j ;

[0053] Where MSF is the position of the fourth stage of the Mountain Gazelle optimization algorithm, Cof 4Searching for performance parameters for the improved algorithm in the fourth phase.

[0054] Furthermore, in step S442, the mathematical model of BH is:

[0055]

[0056] In the formula, X ra is a random mountain gazelle position within the individual interval ra, M pr is the average number of randomly selected mountain gazelle individuals.

[0057] Furthermore, in step S442, the mathematical model of the weight coefficient of the position update strategy of the first stage of the mountain gazelle optimization algorithm is:

[0058]

[0059] Furthermore, in step S444, the mathematical model of the control factor W of the mountain gazelle over the territory is:

[0060] W=(|X(t)|+|X best |)×(2×r 6 -1).

[0061] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0062] In the initialization stage of the mountain gazelle optimization algorithm, the algorithm population is initialized through the good point set strategy, which makes the mountain gazelle optimization algorithm have better traversal, helps the algorithm to obtain the best global optimization effect, and improves the optimization accuracy of the mountain gazelle optimization algorithm; improves the search performance control factor a in a segmented way, so that its value in the early stage remains large, thereby ensuring that the algorithm is mainly global search in the early stage, and the value in the later stage remains small, thereby ensuring that the algorithm is mainly convergent in the later stage, so that the global search and local development are better balanced; "change convergence" mechanism, after the mountain gazelle optimization algorithm position update is completed, when the mountain gazelle optimization algorithm falls into the local optimal, the "change convergence" mechanism is used to jump out of the local optimal. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 Flowchart of the optimization method for laparoscopic image enhancement.

[0064] Figure 2 This is a fitness comparison chart between the improved mountain gazelle optimization algorithm and the standard mountain gazelle optimization algorithm.

[0065] Figure 3 The effect diagram of initializing the algorithm population for the good point set strategy and the random strategy.

[0066] Figure 4Schematic diagram of the comparison between the enhanced abdominal image and the original abdominal image. DETAILED DESCRIPTION

[0067] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0068] See also Figure 1-Figure 4 , the present invention provides a technical solution:

[0069] A laparoscopic image enhancement optimization method is used to enhance the brightness of the abdominal image through the CLAHE algorithm, and then the enhanced abdominal image is segmented by optimizing the K-means algorithm through the improved snow melting optimization algorithm. Figure 1 As shown, the specific steps are as follows.

[0070] S1. In the initialization stage of the mountain gazelle optimization algorithm, the algorithm population is initialized through the good point set strategy, and the dolphin position update strategy in the initialization stage of the algorithm is improved.

[0071] S2. Improve the search performance control factor a of the Mountain Gazelle optimization algorithm by a segmented method, and improve the algorithm search performance parameter Cof i The mathematical model of .

[0072] S3. Introduce the "variable convergence" mechanism in the convergence process of the mountain gazelle optimization algorithm to improve the optimization mechanism of the mountain gazelle optimization algorithm.

[0073] S4. Use the improved mountain gazelle optimization algorithm (IMGO) to adjust the correction coefficient α of the MSRCR algorithm and establish the IMGO-MSRCR algorithm model.

[0074] S5. Input the salient point set of the abdominal cavity image into the IMGO-MSRCR algorithm model, and perform contrast enhancement processing on the abdominal cavity image according to the information of the salient points.

[0075] Preferably, the position update strategy of the original initialization phase of the mountain gazelle optimization algorithm is implemented by a random initialization method, such as Figure 3 As shown in the figure, random initialization will make the population individuals of the mountain gazelle optimization algorithm very concentrated in some areas and very dispersed in other areas, which will increase the probability of the algorithm falling into the local optimal solution too early in the search process, and reduce the stability of the mountain gazelle optimization algorithm; while the initialization of the good point set is more evenly distributed than random nodes, which makes the mountain gazelle optimization algorithm have better traversal, which is conducive to the algorithm to obtain the best global optimization effect and improve the optimization accuracy of the mountain gazelle optimization algorithm.

[0076] Furthermore, in the initial stage of the mountain gazelle optimization algorithm, the initialized mountain gazelle represents the candidate solution of the problem to be optimized, and the position of the mountain gazelle represents the problem variable, that is, the correction coefficient α solution set of the MSRCR algorithm. The correction coefficient α solution set of the MSRCR algorithm is encoded into the individual position of the mountain gazelle updated in the optimization process of the mountain gazelle optimization algorithm through the real number coding method. The mathematical model formula is:

[0077]

[0078] In the formula, is the initial position value of the j-th dimension of the i-th mountain gazelle, lb j is the lower limit of the position of the j-th dimension, ub j is the upper limit of the position of the j-th dimension, P N (i) is a set of good points with the number N, where N is the population size of the mountain gazelle optimization algorithm.

[0079] Preferably, the search performance control factor a of the Mountain Gazelle optimization algorithm affects the search performance parameter Cor i , which determines the optimization speed of the mountain gazelle optimization algorithm. The search performance control factor a before the improvement adopts a linear change method to change with the number of iterations. Regardless of the early or late algorithm iterations, the same global search speed and local development speed are maintained, resulting in the global search and local development performance of the mountain gazelle optimization algorithm cannot be well balanced; the search performance control factor a is improved in a piecewise manner, so that its value in the early stage remains large, thereby ensuring that the algorithm is mainly based on global search in the early stage, and the value in the later stage remains small, thereby ensuring that the algorithm is mainly based on convergence in the later stage, so that the global search and local development are better balanced.

[0080] Furthermore, the improved search performance control factor a 2 The mathematical model formula is:

[0081]

[0082] In the formula, a max is the maximum value of the search performance control factor, a min Minimum value of the search performance control factor.

[0083] Preferably, a "variable convergence" mechanism is introduced in the convergence process of the mountain gazelle optimization algorithm. After the position update of the mountain gazelle optimization algorithm is completed, when the mountain gazelle optimization algorithm falls into a local optimum, the "variable convergence" mechanism is used to jump out of the local optimum, thereby improving the optimization accuracy of the mountain gazelle optimization algorithm.

[0084] Preferably, the mathematical model formula of the MSRCR algorithm is:

[0085]

[0086] In the formula, R MSRCR is the enhanced abdominal image, c i (x, y) is the color restoration factor of the i-th pixel of the original abdominal image, R is the total number of pixels, α is the correction coefficient of the MSRCR algorithm, S i (x, y) is the original pixel image of the abdominal cavity, F i (x, y) is the abdominal cavity pixel image after filtering.

[0087] Preferably, the original abdominal cavity image is preprocessed by low-pass filtering using a Gaussian function to obtain a filtered abdominal cavity pixel image F i (x ， y).

[0088] Preferably, a significant point set is extracted from the preprocessed abdominal image, and the color restoration factor c is adjusted according to the information of the significant point set; for a region with high significance, the c value is appropriately increased to enhance the contrast of the region, and for a region with low significance, the c value is appropriately reduced to maintain the color stability of the region; the original abdominal image S i (x, y) and the filtered abdominal image F i (x, y)) is converted to the logarithmic domain for difference to obtain the reflection image in the logarithmic domain, and the reflection images of multiple scales are weighted summed. Finally, the enhanced abdominal image is output.

[0089] Preferably, when the MSRCR processes an image, there is color distortion. When the correction coefficient α of the MSRCR algorithm is adjusted by the improved mountain gazelle optimization algorithm (IMGO), the performance of the correction coefficient α of the MSRCR algorithm is reflected according to the contrast and Tenengrad. The fitness value function of the improved mountain gazelle optimization algorithm (IMGO) when adjusting the correction coefficient α of the MSRCR algorithm is designed by taking the sum of the contrast value and the reciprocal of the Tenengrad value as the objective function J. The fitness function is designed by the minimum value of the objective function. The smaller the fitness function value is, the more the correction coefficient α of the MSRCR algorithm is adjusted by the improved mountain gazelle optimization algorithm (IMGO). The requirements of MSRCR in processing abdominal images are more met; the mathematical model formula of the objective function J is:

[0090]

[0091] Where max(σ) is the maximum standard deviation of the abdominal image, μ is the mean of the abdominal image, MN is the total number of pixels in the abdominal image, M is the number of horizontal pixels in the abdominal image, N is the number of vertical pixels in the abdominal image, and x is m (u, v) and x n(u, v) is the difference between the pixels in the horizontal and vertical directions of the abdominal image pixel (u, v).

[0092] Preferably, the specific steps of adjusting the correction coefficient α of the MSRCR algorithm using the improved mountain gazelle optimization algorithm (IMGO) are:

[0093] S41. Improve the mountain gazelle optimization algorithm through good point set strategy initialization, including the population size NP of the mountain gazelle algorithm, the problem dimension D, and the initial position of the mountain gazelle The lower limit of the position of the jth dimension of the algorithm lb j and the j-th dimension position upper limit ub j And the maximum number of iterations T, the maximum value of the search performance control factor a max and search performance control factor minimum value a min ;

[0094] S42, encoding the correction coefficient α of the mathematical model of the MSRCR enhanced abdominal image as the location of the individual mountain gazelle;

[0095] S43, simulate the social behavior of mountain gazelles, establish the algorithm mathematical model in four stages, update the individual positions of mountain gazelle population, and the update of individual positions of mountain gazelles is the update of the correction coefficient α solution of MSRCR;

[0096] S44, calculating the fitness value of the individual position of the mountain gazelle population after this iteration update through the objective function, and retaining the individual position of the mountain gazelle population corresponding to the minimum fitness value;

[0097] S45, compare the minimum fitness value of the current iteration with the fitness value of the previous iteration, calculate the speed reduction ratio q of the fitness values ​​before and after, when q≤0.27, it means that the algorithm optimization falls into the global suboptimal solution, use the "change convergence" mechanism to change the neighborhood range, and update the population position to jump out of the global suboptimal solution;

[0098] S46. Use the greedy selection strategy to retain the individual position of the mountain gazelle population with the smallest global fitness value. The current iteration t=t+1. Determine whether the current iteration t is greater than the total number of iterations T. If so, end the algorithm optimization and output the global optimal individual position of the mountain gazelle population. After analysis, assign it to the correction coefficient α of the mathematical model of MSRCR enhanced abdominal images.

[0099] Furthermore, in said S43, the social behavior of mountain gazelles is simulated, an algorithm mathematical model is established in four stages, and the individual positions of the mountain gazelle population are updated, and the specific steps are as follows:

[0100] S431, the improved algorithm searches for performance parameters Cof i After segmentation, the mathematical model was applied to the four stages of mountain gazelle social behavior, and the improved algorithm search performance parameter Cofi The mathematical model formula is:

[0101]

[0102] In the formula, r 3 and r 4 is a random number between 0 and 1, N 2 (D), N 3 (D) and N 4 (D) is a random number within the dimension D of the algorithm problem;

[0103] S442, executing the first phase position update strategy of the improved mountain gazelle optimization algorithm (IMGO), and updating the individual positions of the mountain gazelle population according to formula (1):

[0104] TSM=X best -|(r 1 ×BH-r 2 ×X(t))×F|×Cof 1 (1);

[0105] Where TSM is the position of the first stage of the Mountain Gazelle optimization algorithm, X best is the global optimal position of the current iteration, R 1 and r 2 is a random number between 0 and 1, BH is the individual coefficient that has the ability to reach the optimal position, X(t) is the position of the mountain gazelle in the current iteration, F is the weight coefficient of the position update strategy in the first stage of the mountain gazelle optimization algorithm, Cof 1 Searching for performance parameters for the improved algorithm of the first phase;

[0106] S443, executing the position update strategy of the second phase of the improved mountain gazelle optimization algorithm (IMGO), and updating the individual positions of the mountain gazelle population according to formula (2):

[0107] MH=(BH+Cof 2 )+(r 3 ×X best -r 4 ×X rand )×Cof 2 ;

[0108] Where MH is the position of the second stage of the Mountain Gazelle optimization algorithm, Cof 2 Search performance parameters for the improved algorithm in the second phase, X rand is the vector position of a gazelle randomly selected from the entire population;

[0109] S444, executing the position update strategy of the third phase of the improved mountain gazelle optimization algorithm (IMGO), and updating the individual positions of the mountain gazelle population according to formula (3):

[0110] BMH=(X(t)-W)+(r 5 ×X best -r 6 ×BH)×Cof 3 ;

[0111] Where BMH is the position of the mountain gazelle in the third stage of the optimization algorithm, W is the control factor of the mountain gazelle on the territory, and r 5 and r 6 is a random number between 0 and 1, Cof 3 Searching for performance parameters for the improved algorithm in the third phase;

[0112] S445, executing the fourth stage position update strategy of the improved mountain gazelle optimization algorithm (IMGO), and updating the individual positions of the mountain gazelle population according to formula (4):

[0113] MSF=(ub j- lb j )×Cof 4 +lb j ;

[0114] Where MSF is the position of the fourth stage of the Mountain Gazelle optimization algorithm, Cof 4 Searching for performance parameters for the improved algorithm in the fourth phase.

[0115] Furthermore, in step S442, the mathematical model of BH is:

[0116]

[0117] In the formula, X ra is a random mountain gazelle position within the individual interval ra, M pr is the average number of randomly selected mountain gazelle individuals.

[0118] Furthermore, in step S442, the mathematical model of the weight coefficient of the position update strategy of the first stage of the mountain gazelle optimization algorithm is:

[0119]

[0120] Furthermore, in step S444, the mathematical model of the control factor W of the mountain gazelle over the territory is:

[0121] W=(|X(t)|+|X best |)×(2×r 6 -1).

[0122] Furthermore, the improved mountain gazelle optimization algorithm (IMGO) mathematical model and the MSRCR algorithm were implemented in Matlab through program editing. The original abdominal image data were input into Matlab, and the MSRCR algorithm was optimized by the improved mountain gazelle optimization algorithm (IMGO) to enhance the effect of abdominal image. The population size of the mountain gazelle algorithm was set to NP = 100, the problem dimension was D = 1, and the initial position of the mountain gazelle was set to The lower limit of the position of the jth dimension of the algorithm lb j = -5 and the upper limit of the position of the j-th dimension ub j = 190 and the maximum number of iterations T = 30, the maximum value of the search performance control factor a max = 1.2 and the minimum value of the search performance control factor a min =0.2.

[0123] Furthermore, if Figure 2 As shown in the figure, the fitness value obtained by adjusting the correction coefficient α of the MSRCR algorithm by IMGO is compared with the fitness value obtained by adjusting the correction coefficient α of the MSRCR algorithm by MGO. From the comparison in the figure, IMGO has a smaller fitness value than MGO in the early stage of iteration, indicating that IMGO has found a better correction coefficient α of the MSRCR algorithm from the beginning; IMGO reaches the minimum fitness faster than MGO, indicating that the optimization speed of the MGO algorithm is faster than that of the MGO algorithm, and the fitness value of the IMGO algorithm is smaller than that of the MGO algorithm, indicating that the IMGO algorithm optimizes the correction coefficient α of the MSRCR algorithm with better accuracy.

[0124] Furthermore, the abdominal image is enhanced using the MSRCR algorithm optimized by the IMGO algorithm. The enhanced image is compared with the original abdominal image. Figure 4 As shown, the enhanced abdominal images have more obvious details, higher contrast, and better image display effects for important parts.

Claims

1. A method for laparoscopic image enhancement optimization, characterized in that: The specific steps include: S1. Initialize the algorithm population through the good point set strategy in the initialization stage of the mountain gazelle optimization algorithm, and improve the mountain gazelle position update strategy in the initialization stage of the algorithm; the mathematical model formula is: ； In the formula, For the Mountain Gazelle The initial position value of the dimension, For the The lower limit of the dimension position, For the The upper limit of the dimension position, is a set of good points with the number N, where N is the population size of the mountain gazelle optimization algorithm; S2. Improve the search performance control factor of the Mountain Gazelle optimization algorithm through a segmented approach , Improved algorithm search performance parameters Mathematical model of the improved search performance control factor The mathematical model formula is: ； In the formula, is the maximum value of the search performance control factor, which is 1.

2. is the minimum value of the search performance control factor, which is 0.2, t is the current iteration, and T is the total number of iterations; S3. Introduce the "variable convergence" mechanism in the convergence process of the mountain gazelle optimization algorithm to improve the optimization mechanism of the mountain gazelle optimization algorithm; S4. Use the Improved Mountain Gazelle Optimization (IMGO) algorithm to adjust the correction coefficient of the MSRCR algorithm , establish the IMGO-MSRCR algorithm model, where the mathematical model formula of the MSRCR algorithm is: ； In the formula, For enhanced abdominal images, The original image of the abdominal cavity The color restoration factor of each pixel is is the total number of pixels, is the correction coefficient of the MSRCR algorithm, is the original pixel image of the abdominal cavity, is the abdominal cavity pixel image after filtering; the specific steps of S4 are: S41. Improve the mountain gazelle optimization algorithm through good point set strategy initialization, including the population size NP of the mountain gazelle algorithm and the problem dimension , Mountain Gazelle initial position 、Algorithm No. The lower limit of the dimension's position and Upper limit of dimension position and the maximum number of iterations , search for the maximum value of the performance control factor and search performance control factor minimum ; S42, the correction coefficient of the mathematical model of MSRCR enhanced abdominal images coded as the location of individual mountain gazelles; S43, simulate the social behavior of mountain gazelles, establish an algorithm mathematical model in four stages, update the individual positions of mountain gazelle populations, and the update of individual positions of mountain gazelles is the correction coefficient of MSRCR Solution update; S44, calculating the fitness value of the individual position of the mountain gazelle population after this iteration through the objective function, and retaining the individual position of the mountain gazelle population corresponding to the minimum fitness value; wherein, the sum of the contrast value and the inverse of the Tenengrad value is used as the objective function Design of the Improved Mountain Gazelle Optimization Algorithm (IMGO) to Tune the Correction Factor of the MSRCR Algorithm The fitness value function when The mathematical model formula is: ； In the formula, is the maximum standard deviation of the abdominal image, is the mean value of the abdominal image, is the total number of pixels in the abdominal cavity image, is the number of horizontal pixels of the abdominal image, is the number of longitudinal pixels of the abdominal image, and The abdominal image pixels are The difference between pixels in horizontal and vertical directions; S45, compare the minimum fitness value of the current iteration with the fitness value of the previous iteration, and calculate the speed reduction ratio of the two fitness values ​​before and after ,when When , it means that the algorithm falls into the global suboptimal solution. The "change convergence" mechanism is used to change the neighborhood range and update the population position to jump out of the global suboptimal solution. S46. Use the greedy selection strategy to retain the individual position of the mountain gazelle population with the smallest global fitness value. The current iteration t=t+1. Determine whether the current iteration t is greater than the total number of iterations T. If so, end the algorithm optimization and output the global optimal individual position of the mountain gazelle population. After analysis, assign it to the correction coefficient of the mathematical model of MSRCR enhanced abdominal images. ; S5. Input the salient point set of the abdominal cavity image into the IMGO-MSRCR algorithm model, and perform contrast enhancement processing on the abdominal cavity image according to the information of the salient points.

2. A method for laparoscopic image enhancement optimization according to claim 1, characterized in that: In S3, after the position update of the Mountain Gazelle Optimization Algorithm is completed, when the Mountain Gazelle Optimization Algorithm falls into a local optimum, the "variation convergence" mechanism is used to jump out of the local optimum. Specifically, the variable neighborhood search algorithm is integrated. When a solution better than the current solution is not found in the current neighborhood search, the Mountain Gazelle Optimization Algorithm jumps to the next neighborhood to continue searching; when a solution better than the current solution is found in the current neighborhood search, it jumps back to the first neighborhood to restart the search.

3. The method for laparoscopic image enhancement optimization according to claim 2, characterized in that: In S43, the social behavior of mountain gazelles is simulated, and an algorithm mathematical model is established in four stages to update the individual positions of the mountain gazelle population. The specific steps are: S431, the improved algorithm search performance parameters After segmentation, the mathematical model was applied to the four stages of mountain gazelle social behavior, and the improved algorithm searched for performance parameters The mathematical model formula is: ； In the formula, and is a random number between 0 and 1. , as well as is a random number within the dimension D of the algorithm problem; S432, executing the first phase position update strategy of the improved mountain gazelle optimization algorithm (IMGO), and updating the individual positions of the mountain gazelle population according to formula (1): （1）； In the formula, The position of the first stage of the Mountain Gazelle optimization algorithm, is the global optimal position of the current iteration, and is a random number between 0 and 1. is the individual coefficient that has to reach the optimal position, is the position of the mountain gazelle in the current iteration, is the weight coefficient of the position update strategy in the first stage of the mountain gazelle optimization algorithm, Searching for performance parameters for the improved algorithm of the first phase; S433, executing the position update strategy of the second phase of the improved mountain gazelle optimization algorithm (IMGO), and updating the individual positions of the mountain gazelle population according to formula (2): ； In the formula, The position of the second stage of the Mountain Gazelle optimization algorithm, Searching for performance parameters for the improved algorithm in the second phase, is the vector position of a gazelle randomly selected from the entire population; S434, executing the position update strategy of the third phase of the improved mountain gazelle optimization algorithm (IMGO), and updating the individual positions of the mountain gazelle population according to formula (3): ； In the formula, The position of the third stage of the Mountain Gazelle optimization algorithm, The factors that control the territory of mountain gazelles. and is a random number between 0 and 1. Searching for performance parameters for the improved algorithm in the third phase; S435, executing the fourth stage position update strategy of the improved mountain gazelle optimization algorithm (IMGO), and updating the individual positions of the mountain gazelle population according to formula (4): ； In the formula, The position of the fourth stage of the Mountain Gazelle optimization algorithm, Searching for performance parameters for the improved algorithm in the fourth phase.

Images