LSM-YOLO: A Compact and Effective ROI Detector for Medical Detection

Zhongwen Yu,Qiu Guan,Jianmin Yang,Zhiqiang Yang,Qianwei Zhou,Yang Chen,Feng Chen

2024-08-26

Abstract:In existing medical Region of Interest (ROI) detection, there lacks an algorithm that can simultaneously satisfy both real-time performance and accuracy, not meeting the growing demand for automatic detection in medicine. Although the basic YOLO framework ensures real-time detection due to its fast speed, it still faces challenges in maintaining precision concurrently. To alleviate the above problems, we propose a novel model named Lightweight Shunt Matching-YOLO (LSM-YOLO), with Lightweight Adaptive Extraction (LAE) and Multipath Shunt Feature Matching (MSFM). Firstly, by using LAE to refine feature extraction, the model can obtain more contextual information and high-resolution details from multiscale feature maps, thereby extracting detailed features of ROI in medical images while reducing the influence of noise. Secondly, MSFM is utilized to further refine the fusion of high-level semantic features and low-level visual features, enabling better fusion between ROI features and neighboring features, thereby improving the detection rate for better diagnostic assistance. Experimental results demonstrate that LSM-YOLO achieves 48.6% AP on a private dataset of pancreatic tumors, 65.1% AP on the BCCD blood cell detection public dataset, and 73.0% AP on the Br35h brain tumor detection public dataset. Our model achieves state-of-the-art performance with minimal parameter cost on the above three datasets. The source codes are at: <a class="link-external link-https" href="https://github.com/VincentYuuuuuu/LSM-YOLO" rel="external noopener nofollow">this https URL</a>.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The paper attempts to address the issue of the lack of algorithms that can simultaneously meet real-time performance and high accuracy in Region of Interest (ROI) detection in medical images. Specifically, although the existing basic YOLO framework ensures real-time detection speed, it still faces challenges in maintaining accuracy. Therefore, the authors propose a new model—Lightweight Stream Matching YOLO (LSM-YOLO), which combines Lightweight Adaptive Extraction (LAE) and Multi-Path Stream Feature Matching (MSFM). This model aims to improve the detection accuracy and efficiency of ROI in medical images, reduce the impact of noise, and better integrate high-level semantic features with low-level visual features, thereby enhancing diagnostic assistance capabilities. Experimental results show that LSM-YOLO achieves excellent performance on multiple datasets, including pancreatic tumor, blood cell detection, and brain tumor detection.