Visual Object Tracking on LaSOT 1.0 (test)
74.2AUCLORAT-L
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| LORAT-LBackbone=Large, Search Region=224, Protocol=Supervised2024.11 | 74.2 | 83.6 | 80.9 | — | |
| SAMURAI-LBackbone=Large, Protocol=Zero-shot2024.11 | 74.2 | 82.7 | 80.2 | — | |
| ODTrack-LBackbone=Large, Search Region=384, Protocol=Supervised2024.11 | 74 | 84.2 | 82.3 | — | |
| ODTrack-BBackbone=Base, Search Region=384, Protocol=Supervised2024.11 | 73.2 | 83.2 | 80.6 | — | |
| EVPTrack-BBackbone=Base, Search Region=384, Protocol=Supervised2024.11 | 72.7 | 82.9 | 80.3 | — | |
| HIPTrack-BBackbone=Base, Search Region=384, Protocol=Supervised2024.11 | 72.7 | 82.9 | 79.5 | — | |
| AQATrack-LBackbone=Large, Search Region=384, Protocol=Supervised2024.11 | 72.7 | 82.9 | 80.2 | — | |
| LORAT-BBackbone=Base, Search Region=224, Protocol=Supervised2024.11 | 71.7 | 80.9 | 77.3 | — | |
| SeqTrack-BBackbone=Base, Search Region=384, Protocol=Supervised2024.11 | 71.5 | 81.1 | 77.8 | — | |
| SwinTrack-BBackbone=Base, Search Region=384, Protocol=Supervised2024.11 | 71.4 | 79.4 | 76.5 | — | |
| AQATrack-BBackbone=Base, Search Region=256, Protocol=Supervised2024.11 | 71.4 | 81.9 | 78.6 | — | |
| OSTrackSearch Region=384, Protocol=Supervised2024.11 | 71.1 | 81.1 | 77.6 | — | |
| ARTrack-BBackbone=Base, Search Region=256, Protocol=Supervised2024.11 | 70.8 | 79.5 | 76.2 | — | |
| SAMURAI-BBackbone=Base, Protocol=Zero-shot2024.11 | 70.7 | 78.7 | 76.2 | — | |
| TaMOs-BBackbone=Base, Search Region=384, Protocol=Supervised2024.11 | 70.2 | 79.3 | 77.8 | — | |
| SAMURAI-SBackbone=Small, Protocol=Zero-shot2024.11 | 70 | 77.6 | 75.2 | — | |
| GRM-BBackbone=Base, Search Region=256, Protocol=Supervised2024.11 | 69.9 | 79.3 | 75.8 | — | |
| ROMTrack-BBackbone=Base, Search Region=256, Protocol=Supervised2024.11 | 69.3 | 78.8 | 75.6 | — | |
| SAMURAI-TBackbone=Tiny, Protocol=Zero-shot2024.11 | 69.3 | 76.4 | 73.8 | — | |
| MixFormerSearch Region=288, Protocol=Supervised2024.11 | 69.2 | 78.7 | 74.7 | — | |
| STARKSearch Region=320, Protocol=Supervised2024.11 | 67.1 | 76.9 | 72.2 | — | |
| TransTSource=Ours2021.03 | 64.9 | 73.8 | 69 | — | |
| TransTSearch Region=256, Protocol=Supervised2024.11 | 64.9 | 73.8 | 69 | — | |
| SiamR-CNNSource=CVPR20202021.03 | 64.8 | 72.2 | — | — | |
| PrDiMPSource=CVPR20202021.03 | 59.8 | 68.8 | 60.8 | — | |
| AutoMatchSearch Region=255, Protocol=Supervised2024.11 | 58.2 | 67.5 | 59.9 | — | |
| DiMPSource=ICCV20192021.03 | 56.9 | 65 | 56.7 | — | |
| DiMPSearch Region=288, Protocol=Supervised2024.11 | 56.3 | 64.1 | 56 | — | |
| OceanSource=ECCV20202021.03 | 56 | 65.1 | 56.6 | — | |
| SiamAttnSource=CVPR20202021.03 | 56 | 64.8 | — | — | |
| KYSSource=ECCV20202021.03 | 55.4 | 63.3 | — | — | |
| SiamFC++Source=AAAI20202021.03 | 54.4 | 62.3 | 54.7 | — | |
| MAMLSource=CVPR20202021.03 | 52.3 | — | — | — | |
| CGACDSource=CVPR20202021.03 | 51.8 | 62.6 | — | — | |
| ATOMSource=CVPR20192021.03 | 51.5 | 57.6 | 50.5 | — | |
| SiamBANSource=CVPR20202021.03 | 51.4 | 59.8 | 52.1 | — | |
| SiamCARSource=CVPR20202021.03 | 50.7 | 60 | 51 | — | |
| SiamRPN++Source=CVPR20192021.03 | 49.6 | 56.9 | 49.1 | — | |
| SiamRPN++Protocol=Supervised2024.11 | 49.6 | 56.9 | 49.1 | — | |
| MDNetSource=CVPR20162021.03 | 39.7 | 46 | 37.3 | — | |
| SiamFCSource=ECCVW20162021.03 | 33.6 | 42 | 33.9 | — | |
| ECOSource=ICCV20172021.03 | 32.4 | 33.8 | 30.1 | — | |
| SwinTrack-T-224 (Absolute sine position encoding)Speed (fps)=103, Params (M)=21.6, Positional Encoding=Absolute sine2021.12 | — | — | — | 65.7 | |
| SwinTrack-T-224 (Baseline)Speed (fps)=98, Params (M)=22.7, Backbone=Swin-Transformer, Fusion=Concatenation, Decoder=Target-aware, Positional Encoding=United, Loss=IoU-aware, Hanning penalty=with2021.12 | — | — | — | 66.7 | |
| SwinTrack-T-224 (Cross-attention fusion)Speed (fps)=72, Params (M)=34.6, Fusion=Cross attention-based2021.12 | — | — | — | 66.6 | |
| SwinTrack-T-224 (Plain binary cross entropy loss)Speed (fps)=98, Params (M)=22.7, Loss=Plain binary cross entropy2021.12 | — | — | — | 66.2 | |
| SwinTrack-T-224 (Removing Hanning penalty)Speed (fps)=98, Params (M)=22.7, Hanning penalty=without2021.12 | — | — | — | 65.7 | |
| SwinTrack-T-224 (ResNet-50 backbone)Speed (fps)=121, Params (M)=20.0, Backbone=ResNet-502021.12 | — | — | — | 64.2 | |
| SwinTrack-T-224 (Target query-based decoder)Speed (fps)=91, Params (M)=25.3, Decoder=Target query-based2021.12 | — | — | — | 66.6 |