| 基于改进YOLOv8的脑肿瘤图像检测算法 |
 点此下载全文 |
| 引用本文：郑泽毅1，曹嘉璇1，王家琪1，邹北骥2，郭纯3，刘青萍1.基于改进YOLOv8的脑肿瘤图像检测算法[J].计算技术与自动化,2026,(1):18-25 |
| 摘要点击次数: 319 |
| 全文下载次数: 251 |
|
|
| 中文摘要:脑肿瘤是严重威胁人类健康的疾病，早期检测对提高治疗效果至关重要。针对现有脑肿瘤检测算法在复杂背景或边缘模糊的情况下检测精度不足的问题，本文提出了一种基于改进YOLOv8的脑肿瘤检测算法。该算法引入协调注意力机制，强化肿瘤区域的特征聚焦；在主干网络中采用深度可分离卷积，降低计算复杂度并提升特征提取效率；结合Bottleneck Transformer模块，增强了全局信息建模能力。实验在Brain Tumor Detection数据集上进行，结果显示检测精度达到93%，相较于原算法提升1.1%，mAP0.5和mAP0.5∶0.95分别提升了2.1%和1.7%。实验结果表明，改进算法在脑肿瘤检测任务中表现出显著优势，为医学影像辅助诊断提供了更加精准和高效的支持。 |
| 中文关键词:脑肿瘤检测 YOLOv8 协调注意力机制 深度可分离卷积 Bottleneck Transformer |
| |
| Brain Tumor Image Detection Algorithm Based on Improved YOLOv8 |
|
|
| Abstract:Brain tumors are a severe threat to human health, and early detection is crucial for improving treatment outcomes. To address the issue of insufficient detection accuracy in existing brain tumor detection algorithms under complex backgrounds or with blurred tumor boundaries, this paper proposes an improved YOLOv8-based brain tumor detection algorithm. The algorithm introduces a coordinate attention mechanism to enhance feature focus on tumor regions; employs depthwise separable convolution in the backbone network to reduce computational complexity and improve feature extraction efficiency; and incorporates the bottleneck transformer module to strengthen global information modeling. Experiments were conducted on the brain tumor detection dataset, and the results show that the detection accuracy reached 93%, an improvement of 1.1% over the original algorithm. mAP0.5 and mAP0.5∶0.95 were increased by 2.1% and 1.7%, respectively. The experimental results demonstrate that the improved algorithm exhibits significant advantages in brain tumor detection tasks, providing more accurate and efficient support for medical image-assisted diagnosis. |
| keywords:brain tumor detection YOLOv8 coordinate attention depthwise separable convolution Bottleneck Transformer |
| 查看全文 查看/发表评论 下载pdf阅读器 |
位访问者,今日访问
次