Table 5 Comparison between the performance of yolov5s trained using \({\mathcal{L}}_{CIoU}\) loss and proposed loss. The results are reported \({\mathcal{L}}_{IoAverage}\) on the test dataset of vld100k-61
From: Vehicle logo detection using an IoAverage loss on dataset VLD100K-61
Loss/evaluation | P | R | mAP0.5 | mAP0.5:0.95 |
|---|---|---|---|---|
\({\mathcal{L}}_{CIoU}\) | 0.992 | 0.985 | 0.991 | 0.753 |
\({\mathcal{L}}_{IoAverage}\) Relative improv. % | 0.993 0.1% | 0.983 − 0.2% | 0.992 0.1% | 0.868 15.27% |