Fashion-AI is a PyTorch code base that implements various sate-of-the-art algorithms related to fashion AI, e.g., compatibility prediction, outfit recommendation, etc.
See INSTALL.md
[1] Context-Aware Visual Compatibility Prediction
CVPR, 2019, [pdf]
[2] Learning Binary Code for Personalized Fashion Recommendation.
CVPR, 2019, [pdf]
[3] DeepFashion2: A Versatile Benchmark for Detection, Pose Estimation, Segmentation and Re-Identification of Clothing Images.
CVPR, 2019, [pdf]
- Compatibility Prediction / Outfit Recommendation
Paper [1] - Visual Retrieval / Personalized Fashion Recommendation
Paper [2] - Fashion Image Object Detection/Classification/Parsing/Segmentation/Attribute Manipulation/Landmark Detection
Paper [3]
| Method | Fill-In-The-Blank (FITB) Accuracy | Compatibility AUC | ||
|---|---|---|---|---|
| Orig.(Easy) | Res.(Hard) | Orig.(Easy) | Res.(Hard) | |
| Siamese Net | 54.2 | 54.4 | 0.85 | 0.85 |
| Bi-LSTM | 68.6 | 64.9 | 0.90 | 0.94 |
| TA-CSN | 86.1 | 65.0 | 0.98 | 0.93 |
| Context (Orig.) (K=0) | 62.2 | 47.0 | 0.86 | 0.76 |
| Context (Ours) (K=0) | 64.8 | 48.3 | 0.86 | 0.77 |
| Context (Orig.) (K=3) | 95.9 | 90.9 | 0.99 | 0.98 |
| Context (Ours) (K=3) | 95.3 | 91.1 | 0.99 | 0.98 |
| Context (Orig.) (K=15) | 96.9 | 92.7 | 0.99 | 0.99 |
| Context (Ours) (K=15) | 96.4 | 92.7 | 0.99 | 0.99 |
Please refer here for details.