A Real-time Detection and Decision Fusion Method for Continuously Scanning Breast Ultrasound Images

Abstract

The invention discloses a real-time detection and decision-making fusion method for continuously scanning breast ultrasound images, and belongs to the technical field of image segmentation and fusion. The real-time detection and decision-making fusion method specifically includes the following steps: first, by building and training a fusion network for automatic segmentation and classification of lesion areas, a corresponding prediction model is obtained; after that, the prediction model is embedded in the cloud for continuous scanning images. Real-time detection of the lesion area in each frame of image; then the accurate detection and classification results of the corresponding cases of the continuous scanning images are obtained through the decision-making algorithm; finally, the detection results obtained by the continuous scanning of the breast area are spatially synthesized by the image fusion and splicing method to generate the overall breast area Three-dimensional prediction of image results, and online feedback of the results to doctors. The method realizes the visual output of the overall three-dimensional prediction image results of the breast region, improves the prediction accuracy and speed of lesion region segmentation and classification, and is convenient for embedding into hardware devices or the cloud.

Description

technical field [0001] The invention relates to a real-time detection and decision fusion method for continuously scanning breast ultrasound images, and belongs to the technical field of image segmentation and fusion. Background technique [0002] In recent decades, breast cancer has become a major public health problem in the current society, and it has become a consensus that breast cancer has become a key screening disease. The clinical imaging diagnosis methods of breast lesions mainly include: X-ray photography, ultrasonography and MRI (magnetic resonance examination). Among them, ultrasonography has no damage to the human body, so it has become the main method for large-scale screening of breast lesions. Breast lesions are divided into benign lesions and malignant lesions, among which benign lesions include mastitis, breast hyperplasia, breast fibroids, breast cysts and other diseases; malignant lesions include breast cancer, ductal carcinoma and other diseases. Diag...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the technical defects of existing breast ultrasound image scanners that cannot detect lesion areas in real time and assist decision-making diagnosis, the present invention proposes a real-time detection and decision-making fusion method for continuously scanning breast ultrasound images, and performs a real-time detection and decision-making fusion method for lesion areas in breast ultrasound images Automatically detect, segment and classify, and obtain accurate detection and classification results through decision-making algorithms, and finally generate the overall three-dimensional prediction image results of the breast area through image fusion and splicing methods

Images