Deep learning colorectal cancer polyp segmentation device based on enhanced multi-scale features

Abstract

The invention discloses a deep learning colorectal cancer polyp segmentation device based on enhanced multi-scale features. The device includes: a data preprocessing module for adjusting and normalizing the resolution of colorectal cancer polyp image data Processing; encoder module, which uses feature extractors with multi-scale residual structures and receptive field block components that capture multi-scale receptive fields to extract the diversity features of polyps; decoder module, which uses dense multi-scale skip connections to transfer context The information realizes the segmentation details, and completes the boundary segmentation through the attention mechanism provided by the local context; and uses the deep supervision technology to calibrate in the upsampling process to alleviate the gradient disappearance or explosion phenomenon during training. The invention solves the problems that small polyps are difficult to distinguish and locate, and the boundaries between polyps and surrounding tissues are ambiguous. At the same time, the introduction of a deep supervision mechanism optimizes the gradient of the network model, accelerates the convergence of the network model, and shortens the training time of the network model. .

Description

technical field [0001] The invention relates to the field of image segmentation based on deep learning, in particular to a deep learning colorectal cancer polyp segmentation device based on enhanced multi-scale features. Background technique [0002] Colorectal cancer (CRC) is one of the most common malignant tumors in the world, and its mortality rate ranks third among all cancers. Studies have shown that most colorectal cancer patients are already in the middle and advanced stages with metastasis when they are found, and 95% of colorectal cancers are caused by colorectal adenomatous polyps, and the whole development process takes about 5-10 years. Excision of the lesions at the adenomatous polyp (polyp) stage can prevent colorectal cancer in time. Therefore, the early detection of polyps is particularly important. [0003] Colonoscopy is considered the best diagnostic tool for early detection and removal of polyps, and is the gold standard for colon cancer screening. Ho...

AI Technical Summary

Problems solved by technology

[0006] The present invention provides a deep learning colorectal cancer polyp segmentation device based on enhanced multi-scale features. The problem of blurred boundaries with surrounding organizations, and the introduction of the deep supervision mechanism optimizes the gradient of the network model, speeds up the convergence of the network model, and shortens the training time of the network model. See the description below for details:

Images