Video Reconstruction on SEED-DV 1 (test)
86.7Semantic Accuracy (2-way, Video)MindCine
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| MindCinenumber of classes=102026.01 | 86.7 | 33.1 | 82.6 | 28.8 | 0.333 | 9.568 | 0.775 | |
| EEG2Videonumber of classes=102026.01 | 85 | 32 | 79.8 | 23 | 0.301 | 8.973 | 0.749 | |
| MindCinenumber of classes=402026.01 | 81.8 | 17.9 | 78.7 | 17.1 | 0.278 | 9.035 | 0.768 | |
| MindCine Text+Imagenumber of classes=40, conditioning=Text+Image2026.01 | 81 | 16.6 | 77.6 | 15.5 | 0.27 | 8.79 | 0.762 | |
| MindCine w/o Perceptionnumber of classes=40, ablation_mode=no perception module2026.01 | 80.8 | 15 | 74.1 | 14.9 | 0.198 | 8.297 | 0.732 | |
| MindCine Text+Depthnumber of classes=40, conditioning=Text+Depth2026.01 | 80.4 | 15.5 | 77.2 | 15 | 0.274 | 8.899 | 0.766 | |
| EEG2Videonumber of classes=402026.01 | 80 | 16.1 | 77.2 | 14.6 | 0.258 | 8.684 | 0.728 | |
| MindCine Text-onlynumber of classes=40, conditioning=Text2026.01 | 79.8 | 15 | 76.6 | 14.2 | 0.269 | 8.683 | 0.753 | |
| MindCine w/o Semanticnumber of classes=40, ablation_mode=no semantic module2026.01 | 79.7 | 13.7 | 73.6 | 11.1 | 0.225 | 7.394 | 0.586 |