融合注意力机制的水果分类检测算法研究OA
Research on fruit classification and detection algorithm integrating attention mechanism
针对水果识别中检测精度相对较低,检测速度较慢等问题,提出基于Triplet 注意力机制的YOLOv11n-Triplet水果分类识别算法,将Triplet注意力机制分别应用到三个检测头前面,增强模型对目标之间的关联和上下文信息的表示能力.实验结果表明,YOLOv11n-Triplet在初始YOLOv11n模型基础上,准确率提升4.1%,平均精度提升1.7%,召回率提升1.7%.为水果分类提供了一定的借鉴意义.
Aiming at the issues of relatively low detection accuracy and slow detection speed in fruit recognition,this paper proposes a fruit classification and detection algorithm named YOLOv11n-Triplet,incorporating the Triplet Attention mechanism.The Triplet At-tention mechanism is applied to the front of each of the three detection heads,enhancing the model's capability to represent correlations between objects and contextual information.Experimental results demonstrate that,compared to the baseline YOLOv11n model,YO-LOv11n-Triplet achieves an increase in accuracy by 4.1%,1.7%improvement in mean average precision(mAP),and 1.7%rise in recall.This provides valuable references for fruit classification research.
杨雪
中通服咨询设计研究院有限公司,江苏 南京 210019
信息技术与安全科学
注意力机制目标识别水果分类YOLO
Attention mechanismObject detectionFruit classificationYOLO
《通信与信息技术》 2026 (5)
45-48,4
评论