Automated recursive separable algorithm for medical image processing

In the modern world, there is an increasing demand for automated information processing systems that reduce the time spent on monotonous human work, increase the efficiency of algorithms and the time it takes to solve assigned tasks. The paper presents the development of an automated recursive-separable digital filter for processing digital medical images. The feature of the automated algorithm is its speed; due to its internal structure, the recursive-separable Laplacian Filter “double pyramid” performs separate processing by row and column of the image, using recirculators. The recirculator itself also affects the speed due to the use of recursion properties. Automation of the algorithm will increase the efficiency of digital image processing, due to automatic enumeration of filter coefficients, which will improve the quality of the original information almost without human intervention. The developed algorithm itself carries out the process of image quality assessment and repeats the digital processing procedure until it reaches the required values of one of the parameters: mean square error (MSE), peak signal-to-noise ratio (PSNR) and structural similarity index of image (SSIM). To test the developed automated algorithm, the software "ALF: Automated LDP Filter" was developed and implemented using QtDesigner and PySide6. The software module consists of three data entry areas, such as the mask size h, the coefficient for lifting A1 and the coefficient for increasing the central element A2, two areas for displaying the image, the original and processed, seven buttons, such as input, output and saving the image, checking the image quality condition, three buttons for calculating the parameters MSE, PSNR and SSIM. As part of the testing, a study was carried out on the performance of the automated recursive separable digital filter using the example of endoscopic images obtained from a robotic surgical complex, which showed the effectiveness of its use, since it was possible to achieve the specified parameter values on the output image.