轻量级多尺度特征融合驱动下的舌苔检测方法研究

打开文本图片集

中图分类号：TP391.4；G434 文献标识码：A 文章编号：1006-8228（2025）08-22-08

Abstract：ThisstudyapliesdeeplearningtotonguecoatingfeaturerecognitionintraditionalChinesemedicine，proposinga lightweightnetworkFEA-YOLOv1lbasedoncomputervision.Thearchitecturerealizesmuli-scalefeaturefusionthroughASPF， FEC3K2，andEC3K2modules，enhancingthedetectioncapabilityfortonguecoatingfeaturesofdiferentshapesandscales，while reducingcomputationalcostsandsuppressingnoiseinterference.Experimentsshowthattheoptimized modelimprovesthemAPby （204号 3.1% andreducesparametersby115，520comparedwiththeoriginalYOLOv11，providinganeficient lightweightsolutionfor automated tongue coating diagnosis.

Keywords：Computer Vision; Multi-scale Feature Fusion; Lightweight Network; Deep Learning

0引言

舌苔作为中医诊断的重要对象，可通过高分辨率系统获取图像，但传统识别依赖医师经验，存在主观性，亟须客观标准化研究。（剩余11475字）