Day45 - Backbone中C3/C2f后添加一层SEAttention在VisDrone上的表现

yolov5n/s/m和yolov8n/s仅在backbone的C3/C2f模块后添加一层SE,在VisDrone上训练200epochs

• Yolov5n-se

  Parameters: 2.666M

  Class	Precision	Recall	mAP50	mAP50-95
  all	0.417	0.324	0.317	0.181
  pedestrian	0.435	0.332	0.336	0.141
  people	0.483	0.237	0.266	0.0945
  bicycle	0.2	0.103	0.0727	0.0267
  car	0.627	0.746	0.75	0.515
  van	0.42	0.389	0.372	0.249
  truck	0.403	0.286	0.26	0.163
  tricycle	0.344	0.228	0.203	0.109
  awning-tricycle	0.251	0.132	0.11	0.0708
  bus	0.556	0.418	0.445	0.296
  motor	0.45	0.372	0.357	0.145

• Yolov5s-se

  Parameters: 9.197M

  Class	Precision	Recall	mAP50	mAP50-95
  all	0.498	0.359	0.377	0.223
  pedestrian	0.542	0.36	0.401	0.179
  people	0.56	0.261	0.312	0.117
  bicycle	0.248	0.144	0.115	0.0477
  car	0.735	0.761	0.789	0.558
  van	0.514	0.415	0.438	0.305
  truck	0.522	0.327	0.354	0.225
  tricycle	0.412	0.279	0.263	0.143
  awning-tricycle	0.305	0.158	0.155	0.0983
  bus	0.633	0.458	0.519	0.374
  motor	0.508	0.428	0.422	0.182

• Yolov5m-se

  Parameters: 25.209M

  Class	Precision	Recall	mAP50	mAP50-95
  all	0.527	0.403	0.416	0.252
  pedestrian	0.562	0.402	0.443	0.208
  people	0.603	0.297	0.353	0.138
  bicycle	0.277	0.179	0.145	0.0624
  car	0.756	0.78	0.805	0.584
  van	0.55	0.442	0.464	0.328
  truck	0.535	0.363	0.401	0.263
  tricycle	0.43	0.324	0.302	0.17
  awning-tricycle	0.342	0.212	0.182	0.112
  bus	0.669	0.562	0.589	0.434
  motor	0.545	0.469	0.471	0.216

• Yolov8n-se

  Parameters: 3.168M

  Class	Precision	Recall	mAP50	mAP50-95
  all	0.431	0.333	0.328	0.191
  pedestrian	0.451	0.352	0.355	0.153
  people	0.493	0.241	0.279	0.1
  bicycle	0.209	0.1	0.0789	0.0297
  car	0.657	0.744	0.755	0.522
  van	0.479	0.363	0.376	0.258
  truck	0.419	0.295	0.289	0.185
  tricycle	0.339	0.245	0.209	0.115
  awning-tricycle	0.257	0.162	0.115	0.0736
  bus	0.563	0.438	0.467	0.32
  motor	0.445	0.387	0.359	0.15

• Yolov8s-se

  Parameters: 11.210M

  Class	Precision	Recall	mAP50	mAP50-95
  all	0.5	0.373	0.385	0.229
  pedestrian	0.548	0.365	0.408	0.184
  people	0.575	0.265	0.322	0.124
  bicycle	0.264	0.141	0.122	0.049
  car	0.743	0.758	0.79	0.565
  van	0.518	0.43	0.436	0.307
  truck	0.522	0.341	0.359	0.235
  tricycle	0.413	0.308	0.279	0.153
  awning-tricycle	0.288	0.16	0.145	0.0885
  bus	0.626	0.514	0.542	0.389
  motor	0.503	0.452	0.442	0.195