Endoscopic Image Enhancement Based on Retinex Theory

In order to solve the practical problems of low brightness, nonuniform illumination and low contrast of electronic endoscopy images, an enhanced algorithm based on Retinex theory was proposed. Considering the characteristics of L1 and L2 norm of image difference and the actual demand of texture suppression for illumination, the optimization expression of Retinex theory was established. Based on variational method, the illumination component and reflection component of image were obtained and the illumination correction and detail stretching were carried out respectively. To improve the contrast of vascular images, the proportion of image color channels was adjusted with an improved logarithmic histogram. Matlab test results show that the algorithm can enhance the image of electronic endoscopy preferably. Compared with the same type of algorithm, the evaluation parameters can be improved with different degrees, showing a certain practical application value. © 2021, Editorial Department of Transaction of Beijing Institute of Technology. All right reserved.