Open Access
Table 3
The performance of multi-class heavy-weighted and tiny YOLOv4, comparing original and models formed by a combination of two single class detectors.
| Model type | Average mAP (%) | Prediction time (ms) |
|---|---|---|
| Combined multi-class heavy-weighted | 96.9 | 104.7 |
| Original multi-class heavy-weighted | 98.8 | 62.9 |
| Combined single class tiny | 91.0 | 52.7 |
| Original single class tiny | 90.1 | 27.8 |
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.