Fiber microendoscope image reconstruction method based on comparative learning and storage medium

• Main Authors: ZHANG HEYE, QU YAWEI, HAN JUNWEI, XU CHENCHU, HAN LONGFEI, CHENG PENGFEI, ZHANG DINGWEN
• Format: Patent
• Language: Chinese; English
• Published: 11.10.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention discloses a fiber microendoscopy image reconstruction method based on comparative learning and a storage medium, and the method comprises the steps: obtaining a batch of optical fiber microendoscopy images, carrying out the image pairing according to whether there is a grid or not, carrying out the pixel filtering of a grid image, and obtaining an approximate-grid image, and carrying out the notch filtering of the grid image, and obtaining a non-grid image; constructing a grid removing model based on a convolution automatic encoder, a convolution automatic decoder, an attention mechanism, gating convolution and the like, and optimizing the model; acquiring an image and inputting the image into the optimization model to generate a clear picture; according to the method, the neural network learns the meshless image, the gridding image and the processed image, the gridding image and the meshless image are combined, and the generated image is closer to the meshless image and is far away from the gri