本文全文图片
本文全文表格
-
拍摄方向
Shooting
direction数据集数量
Number of
datasets不同行为数量
Number of various behaviors采食
Feeding咀嚼
Chewing拱草
Arching前方 Ahead 4 484 5 684 792 1 613 上方 Above 4 320 6 958 960 1 946 -
算法
Algorithm跟踪精度/%
Tracking accuracy鲁棒性/%
RobustnessDeep SORT 58.09 47.28 Staple 57.98 43.69 SiameseRPN 56.14 37.15 -
YOLOv8 BiFormer CoT EMA 精确度/%
Precision召回率/%
Recall平均精确度均值/%
mAP√ 69.76 75.72 75.90 √ √ 70.56 77.56 77.53 √ √ 67.14 81.76 77.82 √ √ 69.75 72.23 73.93 √ √ √ 69.40 85.65 76.64 √ √ √ 67.54 75.80 75.76 √ √ √ 64.32 86.77 77.66 √ √ √ √ 77.73 82.57 83.70 -
YOLOv8 BiFormer CoT EMA 精确度/%
Precision召回率/%
Recall平均精确度均值/%
mAP√ 68.29 77.16 67.23 √ √ 70.93 82.35 72.36 √ √ 73.05 80.52 73.00 √ √ 69.50 80.00 67.63 √ √ √ 73.54 85.92 73.82 √ √ √ 73.24 84.41 72.59 √ √ √ 71.45 82.60 69.47 √ √ √ √ 76.32 86.33 76.81 -
模型
Model精确度
Precision召回率
Recall平均精确度均值
mAP前方
Front上方
Above前方
Front上方
Above前方
Front上方
AboveYOLOv8 71.84 69.97 75.69 76.52 76.33 65.67 BCE-YOLO 79.66 79.23 83.79 82.79 86.94 83.67