System and method for automatically detecting lesions in medical image through multi-model fusion

Abstract

The invention discloses a system and method for automatically detecting lesions in a medical image through multi-model fusion; the method comprises using single or fused detection models including deep learning technology to analyze and detect suspicious lesions in the medical image, such as a breast X-ray image. By using the system and method, lesion characteristics can be automatically extracted; the system and method are suitable for detecting or marking one or more types of lesions.

Description

technical field [0001] The present invention relates to a system and method for automatic detection of medical images combined with deep learning technology, in particular to using a single or fusion detection model including deep learning technology to detect suspicious lesions in medical images (such as mammograms) Systems and methods for detection and evaluation. Background technique [0002] Breast cancer is the most common cancer that threatens women's health. The key to preventing and treating breast cancer is early detection, early diagnosis and early treatment. Common modalities for breast health screening include X-rays, ultrasounds, and magnetic resonance imaging (MRI). Among them, mammography is considered the most accurate detection method, because it can detect various suspicious lesions (such as masses, microcalcifications, structural disorders, etc.) in the early stage. At present, mammogram diagnosis is mainly done by doctors visually, and the quality of d...

AI Technical Summary

Problems solved by technology

[0005] However, using deep learning technology to completely replace traditional computer-aided detection technology also has its shortcomings.

If a single deep learning strategy is used to realize computer-aided diagnosis, there will be a lack of comprehensive consideration of the combined use of various traditional computer-aided detection models and multiple detection models, and it may not be the optimal detection model solution.

Images