Auxiliary diagnosis system for early gastric cancer under electronic staining endoscope based on deep learning

Abstract

The invention discloses an auxiliary diagnosis system for early gastric cancer under an electronic staining endoscope based on deep learning, and the system comprises an image obtaining module which obtains a stomach image collected in real time in an examination process of the electronic staining endoscope; an early gastric cancer recognition module used for recognizing a focal zone and marking the focus area based on a pre-constructed early gastric cancer recognition model; and a logic judgment module used for judging whether the identified focal zone belongs to the same focal zone or not when the identified focal zone is overlapped. By means of a common electronic staining means, the invention provides the method for identifying the gastric early cancer based on an electronic staining endoscopic image, and the real-time performance, accuracy and practicability of detection are greatly improved.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence, and in particular relates to an auxiliary diagnosis system for early gastric cancer under electronic chromoendoscopy based on deep learning. Background technique [0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art. [0003] Gastric cancer is a common malignant tumor. At present, artificial intelligence methods have been introduced into the screening of early gastric cancer. However, due to the existence of gastric inflammation background, there are various mucosal manifestations under the microscope in the early screening of early gastric cancer. There are still many difficulties. For example, early gastric cancer usually shows a slight bulge or depression, accompanied by slight redness; another example, the depth of invasion of gastric cancer is unpredictable, differentiated, ...

AI Technical Summary

Problems solved by technology

[0003] Gastric cancer is a common malignant tumor. At present, artificial intelligence methods have been introduced into the screening of early gastric cancer. However, due to the existence of gastric inflammation background, there are various mucosal manifestations under the microscope in the early screening of early gastric cancer. There are still many difficulties

For example, early gastric cancer usually shows a slight bulge or depression, accompanied by slight redness; another example, the depth of invasion of gastric cancer is unpredictable, differentiated, intramucosal gastric cancer or gastric cancer invading superficial submucosa, and gastric cancer invading the submucosa Stomach cancer can be difficult to differentiate, and there is a big difference in how the two situations are handled

[0004] Gastroscopy images obtained by traditional white-light endoscopy, although relevant studies have confirmed that computer-aided diagnosis can improve the diagnostic rate of ordinary white-light endoscopy, there are still many misdiagnosis rates, because ordinary white-light endoscopy can only evaluate the overall Lesions, unable to observe localized lesions

There are also studies using special imaging detection of early gastric cancer, such as magnifying endoscopy combined with narrow-spectrum imaging, which can be used clinically to distinguish cancerous areas from non-cancerous areas, however, this optical diagnosis requires a lot of professional knowledge and experience, which hinders It is widely used in gastroscopic examination, and it is difficult to meet the real-time requirements of examination

Images