Image detection method, device and equipment for intestinal villus form and damaged area and medium

Abstract

The invention relates to the technical field of image detection, and discloses an image detection method, device, equipment and medium for an intestinal villus form and a damaged region, and the method comprises the steps: employing a pre-trained U-Net model, inputting a pre-processed intestinal pathology staining image for training, extracting multi-scale features through an encoder, and carrying out the recognition of the multi-scale features; the decoder fuses the features and outputs a segmentation mask, and a U-Net model is trained through a loss function; performing size adjustment and normalization on a to-be-measured image, inputting the to-be-measured image into the U-Net model, and outputting a predicted thermodynamic diagram; performing dynamic threshold binarization processing on the thermodynamic diagram to obtain a binarized image; extracting a contour in the binarized image, marking the boundary of the damaged region, mapping the contour to a thermodynamic diagram, and performing superposition display; and generating a visual result of the thermodynamic diagram containing the original intestinal pathology staining image, the thermodynamic diagram and the contour mark, wherein the visual result is used for evaluating the villus form of the intestinal tract and the position and the range of the damaged region. According to the invention, information can be obtained more quickly, and the disease development trend can be predicted.

Images