The invention relates to the technical field of medical
image processing, in particular to a gastric
early cancer auxiliary diagnosis method based on a
deep learning multi-model fusion technology, which comprises the following steps: S1, constructing
multiple models; s2, collecting gastroscope images, forming continuous serialized image frames, identifying the
light source mode of the current
image frame by utilizing the image classification model 1, entering the step S3 to mark the position of a focus by the target detection model 2 when the current
image frame is identified as a white lightmode, and marking the high-risk focus by utilizing the image classification model 3; and when the
image frame is identified as the
dyeing amplification mode, entering the step S4 in which a segment model group can extract the boundary range, the
microvessel form and the micro-tissue structure feature map in the image frame in real time, and outputting whether canceration occurs or not, the credibility and the differentiation type by the decision-making model 7. A plurality of
deep learning models are constructed according to different tasks, a parallel
cascade model fusion technology is adopted, and a full-process intelligent auxiliary diagnosis function is provided in the
stomach early cancer screening process of endoscopists.