Meniscus injury grading method based on mixed attention weakness supervised transfer learning

Abstract

The invention discloses a meniscus injury grading method based on mixed attention weakness supervised transfer learning. The meniscus injury grading method mainly solves the problems that an existing method is large in meniscus injury degree grading difficulty and lacks clinical interpretability. The method comprises the steps of positioning a meniscus area through a target detection network to generate a meniscus area image data set; carrying out two-dimensional histogram equalization enhancement operation on the data; then using a transfer learning pre-training network to extract saliency features of the enhanced image to generate a feature map, and using weak supervision attention learning to generate an attention map of the enhanced image; and finally, through a bilinear attention pooling algorithm, classifying the meniscus injury grades in combination with the feature map and the attention map to obtain a visual meniscus injury grading result. According to the method, the requirement for automatic classification diagnosis of half-month injury clinically is met, the interpretability of deep learning in clinical diagnosis is improved, and the method can be used for image detection and classification of nuclear magnetic resonance lesions.

Description

technical field [0001] The invention belongs to the technical field of image processing, and mainly relates to a human knee meniscus positioning and damage classification method, which can be used for image detection and classification of nuclear magnetic resonance lesions. Background technique [0002] The knee joint is the most complex joint in the human body and plays an important role in weight bearing and multiaxial movement. Therefore, the probability of injury is high, and meniscus injury is the most common type of knee joint injury. Under physiological conditions, the meniscus plays an important role in maintaining the stability of the knee joint, buffering shocks, and lubricating the joints. An injured meniscus can cause knee pain, swelling, and locking of the joint, severely limiting a patient's mobility. [0003] Magnetic resonance imaging (MRI) is the highest resolution imaging method for tissue. Compared with X-ray films and CT examinations, MRI has obvious a...

AI Technical Summary

Problems solved by technology

[0008] The purpose of the present invention is to propose a meniscus injury grading method based on mixed attention weak supervision transfer learning, to solve the problem that the prior art cannot classify the severity of knee meniscus injury, and by analyzing the knee meniscus internal The visual display of the injury situation improves the interpretability and provides a more reliable basis for clinical diagnosis

Images