Scholarlink

Aiming to enhance the precision and stability of silkworm disease recognition under complex backgrounds, this paper constructs a Mask-RCNN_ResNet101 model that incorporates the CBAM attention mechanism. The model boosts the feature extraction capabilities of key regions related to silkworm diseases by integrating the CBAM module, effectively suppresses background interference, and improves contour recognition clarity. The backbone network resorts to ResNet101 to strengthen the overall feature representation capabilities. The experimental results demonstrate that the enhanced model surpasses the baseline model in both recognition accuracy (AP enhanced by 3.7%) and recall (increased by 2.1%), and exhibits good adaptability in complex background and blurred contour scenarios. The findings hold certain reference significance for elevating the level of silkworm disease automatic identification and driving the intelligent development of sericulture.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Mask R-CNN Incorporating Attention Mechanism for Silkworm Disease Recognition in Complex Backgrounds

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Peng Chen,
  • Peng Tang,
  • Tin Tin Ting,
  • Maidin Siti Sarah,
  • Shuaiming Qiu,
  • Zilin Huang

摘要

Aiming to enhance the precision and stability of silkworm disease recognition under complex backgrounds, this paper constructs a Mask-RCNN_ResNet101 model that incorporates the CBAM attention mechanism. The model boosts the feature extraction capabilities of key regions related to silkworm diseases by integrating the CBAM module, effectively suppresses background interference, and improves contour recognition clarity. The backbone network resorts to ResNet101 to strengthen the overall feature representation capabilities. The experimental results demonstrate that the enhanced model surpasses the baseline model in both recognition accuracy (AP enhanced by 3.7%) and recall (increased by 2.1%), and exhibits good adaptability in complex background and blurred contour scenarios. The findings hold certain reference significance for elevating the level of silkworm disease automatic identification and driving the intelligent development of sericulture.