Research map / Label-efficient and robust learning / Recognition and long-tail
Fashion Style in 128 Floats: Joint Ranking and Classification Using Weak Data for Feature Extraction
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Jointly train feature extraction with ranking and classification losses | 128-dimensional fashion features · Hipster Wars | accuracy: 73.5 -> 75.9 | Helped |
| Add the classification loss to ranking-only training (ablation) | 128-dimensional fashion features · Hipster Wars | accuracy: 74.5 -> 75.9 | Helped |
| Add the ranking loss to classification-only training (ablation) | 128-dimensional fashion features · Hipster Wars | accuracy: 73.5 -> 75.9 | Helped |
| Clean the training data before joint training (ablation) | Fashion144k-trained features · Hipster Wars | accuracy: 72.9 -> 75.9 | Helped |
| Jointly train with ranking and classification rather than Siamese loss (ablation) | 128-dimensional fashion features · Hipster Wars | accuracy: 73.3 -> 75.9 | Helped |
| Jointly train feature extraction with ranking and classification losses | 128-dimensional fashion features · Fashion144k | accuracy: 14.6 -> 17.0 | Helped |
| Add the classification loss to Siamese training (ablation) | 128-dimensional fashion features · Fashion144k | accuracy: 13.9 -> 17.0 | Helped |
| Train the proposed architecture instead of using random weights (ablation) | 128-dimensional fashion features · Fashion144k | accuracy: 13.0 -> 17.0 | Helped |
| Use the joint method instead of VGG 16 Layers features | fashionability prediction · Fashion144k | accuracy: 16.6 -> 17.0 | Helped |
Closest papers by meaning
- Consensus Style Centralizing Auto-Encoder for Weak Style Classification (AAAI 2016 · similarity 0.84)
- Learning Attribute and Class-Specific Representation Duet for Fine-Grained Fashion Analysis (CVPR 2023 · similarity 0.84)
- RankMix: Data Augmentation for Weakly Supervised Learning of Classifying Whole Slide Images With Diverse Sizes and Imbalanced Categories (CVPR 2023 · similarity 0.84)
- Dress Fashionably: Learn Fashion Collocation With Deep Mixed-Category Metric Learning (AAAI 2018 · similarity 0.84)
- Weakly Supervised Image Classification Through Noise Regularization (CVPR 2019 · similarity 0.83)
- FashionERN: Enhance-and-Refine Network for Composed Fashion Image Retrieval (AAAI 2024 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.