| 轻型多尺度黑色素瘤目标检测网络模型的建立：基于注意力机制调控 |
| |
| 引用本文： | 钟友闻,车文刚,高盛祥.轻型多尺度黑色素瘤目标检测网络模型的建立：基于注意力机制调控[J].南方医科大学学报,2022,42(11):1662-1671. |
| |
| 作者姓名： | 钟友闻 车文刚 高盛祥 |
| |
| 作者单位： | 昆明理工大学信息工程与自动化学院，云南 昆明 650500；昆明理工大学云南省计算机技术应用重点实验室，云南 昆明 650500；昆明理工大学云南省人工智能重点实验室，云南 昆明 650500 |
| |
| 摘 要： | 目的 提出一种融入坐标注意力和高效通道注意力机制的深度学习目标检测模型AM-YOLO。方法 运用Mosaic图像增强与MixUp混类增强对图像进行预处理，采用One-Stage结构的目标检测模型YOLOv5s，并对该模型的骨干网络与颈部网络进行改进。在该模型的骨干网络中把空间金字塔的最大池化层替换成二维最大池化层，接着将坐标注意力机制和高效通道注意力机制分别融入到YOLOv5s模型的C3模块与该模型的骨干网络中。将改进后的模型与未改进的YOLOv5s模型，YOLOv3模型，YOLOv3-SPP模型，YOLOv3-tiny模型进行相关算法指标的对比实验。结果 融入了坐标注意力和高效通道注意力机制的AM-YOLO模型能够有效提升对黑色素瘤的识别率，同时也减少了模型权重的大小。AM-YOLO模型在准确率，召回率以及平均精度均值上都要明显优于其他模型，并且对于早期和晚期黑色素瘤的平均精度均值分别达92.8%和87.1%。结论 本文采用的深度学习目标检测算法模型能够应用于黑色素瘤目标的识别中。
|
| 关 键 词： | 深度学习 目标检测 注意力机制 YOLOv5s 黑色素瘤 |
A lightweight multiscale target object detection network for melanoma based on attention mechanism manipulation |
| |
| ZHONG Youwen,CHE Wengang,GAO Shengxiang.A lightweight multiscale target object detection network for melanoma based on attention mechanism manipulation[J].Journal of Southern Medical University,2022,42(11):1662-1671. |
| |
| Authors: | ZHONG Youwen CHE Wengang GAO Shengxiang |
| |
| Affiliation: | Faculty of Information Engineering and Automation, Kunming University of Science and Technology, Kunming 650500, China; Yunnan Key Laboratory of Computer Technology Applications, Kunming University of Science and Technology, Kunming 650500, China; Yunnan Key Laboratory of Artificial Intelligence, Kunming University of Science and Technology, Kunming 650500, China |
| |
| Abstract: | Objective To propose a deep learning target detection model AM- YOLO that integrates coordinate attention and efficient attention mechanism. Methods Mosaic image enhancement and MixUp mixed-class enhancement were used for image preprocessing. In the target detection model YOLOv5s with One-Stage structure and modified backbone network and neck network, the maximum pooling layer of the spatial pyramid of the backbone network was replaced with a two-dimensional maximum pooling layer, and the coordinate attention mechanism and the efficient channel attention mechanism were integrated into the C3 module and the backbone network of the model, respectively. The improved model was compared with the unmodified YOLOv5s model, YOLOv3 model, YOLOv3-SPP model, and YOLOv3-tiny model for relevant algorithmic indicators in comparative experiments. Results The AM-YOLO model incorporating coordinate attention and efficient channel attention mechanism effectively improved the accuracy of melanoma recognition with also a reduced size of the model weight. This model showed significantly better performance than other models in terms of precision, recall rate and mean average precision, and its mean average precision for benign and malignant melanoma reached 92.8% and 87.1%, respectively. Conclusion The deep learning-based target object detection algorithm model can be applied in recognition of melanoma targets. |
| |
| Keywords: | deep learning object detection attention mechanism YOLOv5s melanoma |
|
| 点击此处可从《南方医科大学学报》浏览原始摘要信息 |
|
点击此处可从《南方医科大学学报》下载全文 |
