Suspect Face Generation on CelebA ID
66.429AF Match RateID-SD (ours)
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| ID-SD (ours)Text input=true, LQ input=true, Sketch input=true, Backbone=SD-1.52026.05 | 66.429 | 30.8 | 84.78 | 58.029 | 29.142 | 86.432 | 48.91 | 63.628 | 0.8709 | |
| ControlNet-SD (Sketch)Text input=true, LQ input=false, Sketch input=true, Backbone=SD-1.52026.05 | 53.297 | 27.352 | 60.769 | 51.084 | 29.531 | 85.269 | 55.084 | 50.314 | 0.2533 | |
| ID-PA (ours)Text input=true, LQ input=true, Sketch input=true, Backbone=PixArt-α2026.05 | 48.736 | 27.193 | 72.198 | 55.431 | 19.775 | 86.706 | 38.949 | 61.218 | 2.6368 | |
| ControlNet-PA (Sketch)Text input=true, LQ input=false, Sketch input=true, Backbone=PixArt-α2026.05 | 33.681 | 22.839 | 48.407 | 49.015 | 33.399 | 83.324 | 53.158 | 51.061 | 2.6909 | |
| DiffFaceSketchText input=false, LQ input=false, Sketch input=true, Backbone=VQF42026.05 | 32.727 | 23.233 | 52.912 | 50.755 | 40.85 | 84.073 | 33.294 | 59.556 | — | |
| ControlNet-PA (Dual Branch)Text input=true, LQ input=true, Sketch input=true, Backbone=PixArt-α2026.05 | 14.011 | 17.808 | 35.659 | 46.506 | 30.698 | 82.972 | 48.176 | 52.249 | 2.6908 | |
| Conditional Image (Condition Baseline)Text input=false, LQ input=true, Sketch input=true, Backbone=PixArt-α2026.05 | 13.991 | 13.33 | 1.648 | 21.006 | 396.18 | 76.966 | 51.342 | 45.436 | — | |
| ID-PA (zero text)Text input=false, LQ input=true, Sketch input=true, Backbone=PixArt-α2026.05 | 13.736 | 16.24 | 19.67 | 40.349 | 37.819 | 83.431 | 51.052 | 51.879 | — | |
| T2I-Adapter (Dual Branch)Text input=true, LQ input=true, Sketch input=true, Backbone=SD-1.52026.05 | 6.592 | 12.623 | 9.725 | 38.093 | 62.506 | 78.089 | 50.304 | 44.832 | 2.2187 | |
| ControlNet-SD (LQ)Text input=true, LQ input=true, Sketch input=false, Backbone=SD-1.52026.05 | 1.319 | 6.02 | 4.67 | 29.301 | 40.181 | 77.018 | 56.67 | 45.455 | 0.1037 | |
| ControlNet-PA (LQ)Text input=true, LQ input=true, Sketch input=false, Backbone=PixArt-α2026.05 | 1.264 | 8.392 | 6.374 | 34.642 | 33.526 | 78.479 | 54.789 | 42.173 | 2.4893 | |
| OSDFaceText input=false, LQ input=true, Sketch input=false, Backbone=SD-2.12026.05 | 0.951 | 5.114 | 4.341 | 31.677 | 34.237 | 79.913 | 43.225 | 51.352 | — | |
| PixArt-α (Model Baseline)Text input=true, LQ input=false, Sketch input=false, Backbone=PixArt-α2026.05 | 0.604 | 4.378 | 0.22 | 18.969 | 84.384 | 70.46 | 71.84 | 37.398 | 4.5887 |