Machine learning-based endoscopic auxiliary biopsy system and method

Abstract

The invention provides a machine learning-based endoscopic auxiliary biopsy system and method, and belongs to the technical field of endoscopes, and the system comprises an image collection module which is used for obtaining a to-be-detected part video frame collected in real time in an endoscopic examination process; a lesion infiltration depth recognition module used for recognizing a lesion area of the video frame of the to-be-detected part by utilizing the lesion infiltration depth recognition model and carrying out infiltration depth scoring on different differentiation types of the lesion area to obtain mask images with scoring matrixes with different infiltration depths; wherein the lesion infiltration depth recognition model is obtained by training a plurality of groups of data, and each group of data comprises an endoscopic image containing a focus area and labeling information used for labeling different differentiation types of the lesion in the endoscopic image. An endoscopic lesion scoring matrix can be displayed in real time according to the lesion differentiation type and the infiltration depth, so that an endoscopic doctor is assisted in selecting the optimal biopsy part, and the optimal diagnosis and treatment decision is obtained clinically.

Description

technical field [0001] The invention relates to the technical field of endoscopic biopsy, in particular to a system and method for endoscopic assisted biopsy based on machine learning. Background technique [0002] Endoscopic biopsy is currently an indispensable method for the early diagnosis and treatment of digestive system diseases. When suspicious lesions are found during endoscopic examination, biopsy is often required to clarify the nature, differentiation type, and infiltration depth of the lesion, which are related to subsequent treatment decisions. In clinical practice, there are often inconsistencies between preoperative biopsy and postoperative pathology. The reason may be that the differentiation types of different parts of the same lesion are inconsistent, and a certain part of the lesion is too deeply infiltrated. How to choose the best biopsy site and reduce surgical time? Front misjudgment is an urgent problem that needs to be resolved. [0003] Although the...

AI Technical Summary

Problems solved by technology

In clinical practice, there are often inconsistencies between preoperative biopsy and postoperative pathology. The reason may be that the differentiation types of different parts of the same lesion are inconsistent, and a certain part of the lesion is too deeply infiltrated. How to choose the best biopsy site and reduce surgical time? Pre-misjudgment is an urgent problem to be solved

[0003] Although the current convolutional neural network technology can solve the problem of image recognition very well, it is mostly based on the judgment of the overall lesion and cannot meet the needs of accurate biopsy.

Images