An Algorithm for Enhancing Contour Features of Tissue Cell Images Based on Fluorescence in Situ Hybridization

With the widespread application of computer technology in medical image processing, preprocessing fluorescence in situ hybridization (FISH) tissue cell images with complex structures to obtain clear contour features is crucial to ensure the effectiveness of artificial intelligence models. Clear image contour features directly affect the performance of artificial intelligence models. Therefore, in-depth research was conducted on how to enhance the contour feature information of complex FISH tissue cell images, and improvements were made based on existing algorithms. An image contour enhancement method combining image illumination and an improved local binary pattern (LBP) algorithm is proposed. This method uses an illumination enhancement algorithm to improve the overall distribution of image brightness, and then uses an improved local binary pattern algorithm to quickly extract the contour features of the image, and weights the extracted features into the illumination-enhanced image to obtain an image with enhanced contour features. Finally, areas with high energy gradients are identified through piecewise linear functions and heat maps, which effectively filter out areas with low energy gradients. The high-frequency signal areas are the outline areas of FISH tissue cell images. Experimental analysis verified the effectiveness of this method in image enhancement.