Table 5 Recognition accuracy (%) of various datasets with STIP+BSIF-TOP and iDT+BSIF-TOP methods using bag-of-visual-words (BoVW) and fisher vector (FV) encoding
From: Exploiting textures for better action recognition in low-quality videos
| Â | STIP+BSIF-TOP | iDT+BSIF-TOP | ||
|---|---|---|---|---|
Datasets | BoVW | FV | BoVW | FV |
KTH- SD 2 | 88.80 | 89.26 | 93.89 | 92.87 |
KTH- SD 3 | 85.28 | 83.15 | 88.33 | 87.78 |
KTH- SD 4 | 81.67 | 80.19 | 82.41 | 81.02 |
KTH- TD 2 | 88.70 | 89.91 | 95.09 | 94.44 |
KTH- TD 3 | 86.11 | 87.78 | 92.22 | 92.59 |
KTH- TD 4 | 84.54 | 82.96 | 90.00 | 90.28 |
Youtube-LQ | 76.05 | 75.04 | 80.45 | 78.13 |
HMDB-BQ | 32.46 | 33.06 | 37.80 | 40.69 |
HMDB-MQ | 37.14 | 38.51 | 45.96 | 51.62 |