基于轻量化YoloV5s的血癌细胞检测  

Detection of Blood Cancer Cells Based on Lightweight YoloV5s

作　　者：乔冬 何利文 

机构地区：[1]南京邮电大学物联网学院,江苏 南京

出　　处：《软件工程与应用》2024年第2期223-233,共11页Software Engineering and Applications

摘　　要：为了解决人工检测血癌细胞耗费人力,且容易出现漏检、误检的情况,提出一种新型的检测血癌细胞检测算法——YoloV5s-FasterNet。该算法在原本的YoloV5s框架上,通过将主干网络中的C3检测层改进为FasterNet轻量化网络用来提取特征信息,降低整体网络模型的复杂程度,并在主干网络中增加坐标注意力(coordinate attention, CA)机制模型更好的定位和识别目标信息来提高检测精度。通过在血癌数据集上进行大量实验结果表明:与原模型相比,改进的模型平均精度提升2.2%,参数量减少25.7%,检测速度提高15%,实验结果良好,证明该算法对血癌细胞检测具有实用性。In order to solve the problem that manual detection of blood cancer cells is labor-intensive, and it is easy to miss and misdetect, a new detection algorithm for blood cancer cells is proposed- YoloV5s-FasterNet. Based on the original YoloV5s framework, this algorithm improves the C3 detection layer in the backbone network into a FasterNet lightweight network to extract feature information, reduce the complexity of the overall network model, and add coordinate attention to the backbone network. This model can better locate and identify target information to improve detection accuracy. A large number of experimental results on the blood cancer data set show that: compared with the original model, the average accuracy of the improved model is increased by 1.2%, the number of parameters is reduced by 25.7%, and the detection speed is increased by 15%. The experimental results are good, proving that the algorithm is effective in detecting blood cancer cells.

关 键 词：深度学习 血癌细胞检测 YOLOV5s 轻量化网络 注意力机制 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]