Ad-hoc Video Search on TRECVID (TV16) 2016 (test)
17.3infAPSEA + Dual Encoding
Evaluation Results
| Method | Links | |
|---|---|---|
| SEA + Dual Encodingfusion_type=Late Average Fusion2020.11 | 17.3 | |
| SEA + SEAfusion_type=Late Average Fusion, models=SEA({BoW,w2v}) + SEA({BoW,w2v,bi-GRU})2020.11 | 16.6 | |
| Dual Encoding2020.11 | 16.5 | |
| SEAsentence_encoders=BoW, w2v, bi-GRU2020.11 | 16.4 | |
| W2VV++2020.11 | 16.2 | |
| Extended Dual Encoding2020.11 | 15.9 | |
| SEAsentence_encoders=BoW, w2v2020.11 | 15.7 | |
| Dual EncodingBackbone=ResNeXt-ResNet2020.09 | 15.2 | |
| W2VV++Backbone=ResNeXt-ResNet2020.09 | 15.1 | |
| W2VVBackbone=ResNeXt-ResNet2020.09 | 14.9 | |
| VSE++2020.11 | 13.5 | |
| VSE++Backbone=ResNeXt-ResNet2020.09 | 13.5 | |
| VideoStorysentence_encoder=BoW2020.11 | 8.7 | |
| VideoStory2020.09 | 8.7 | |
| CEBackbone=ResNeXt-ResNet2020.09 | 7.4 | |
| Markatopoulou et al.2020.09 | 6.4 | |
| Rank 1context=TRECVID 2016 Finalist2020.11 | 5.4 | |
| Rank 1 finalist2020.09 | 5.4 | |
| Rank 2context=TRECVID 2016 Finalist2020.11 | 5.1 | |
| Rank 2 finalist2020.09 | 5.1 | |
| Rank 3context=TRECVID 2016 Finalist2020.11 | 4 | |
| Rank 3 finalist2020.09 | 4 | |
| LAFF-mltraining_split=MV-test1k2021.12 | 0.222 | |
| LAFFtraining_split=MV-test1k2021.12 | 0.211 | |
| CLIP-FTtraining_split=MV-test1k2021.12 | 0.191 | |
| CLIP2Videotraining_split=MV-test1k2021.12 | 0.176 | |
| Rank 12021.12 | 0.054 | |
| Rank 22021.12 | 0.051 | |
| Rank 32021.12 | 0.04 |