Research map / Label-efficient and robust learning / Recognition and long-tail
CNN-RNN: A Unified Framework for Multi-Label Image Classification
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 |
|---|---|---|---|
| model label co-occurrence with an LSTM recurrent layer (ablation) | CNN-RNN · MS-COCO | MAP@10: 57.2 -> 61.2 | Helped |
| model label co-occurrence with an LSTM recurrent layer (ablation) | CNN-RNN · MS-COCO, k=3 | O-R: 61.3 -> 66.4 | Helped |
| combine the CNN image representation with sequential label prediction and label embeddings | multi-label classification · NUS-WIDE 81 concepts, k=3 | O-F1: 53.9 -> 55.2 | Helped |
| combine the CNN image representation with sequential label prediction and label embeddings | multi-label classification · NUS-WIDE 1000 tags, k=10 | MAP@10: 24.8 -> 26.6 | Helped |
| combine the CNN image representation with sequential label prediction and label embeddings | multi-label classification · MS-COCO, k=3 | O-F1: 63.3 -> 67.8 | Helped |
| use CNN-RNN rather than HCP-1000C | PASCAL VOC 2007 · PASCAL VOC 2007 test | mAP: 81.5 -> 84.0 | Helped |
| use learned or fixed alternative label orders instead of frequency-based ordering (ablation) | CNN-RNN · multi-label classification benchmarks | classification performance | No clear effect |
Closest papers by meaning
- Order-Free RNN With Visual Attention for Multi-Label Classification (AAAI 2018 · similarity 0.93)
- Multi-Label Image Recognition With Graph Convolutional Networks (CVPR 2019 · similarity 0.91)
- Semantic Regularisation for Recurrent Image Annotation (CVPR 2017 · similarity 0.90)
- Learning Spatial Regularization With Image-Level Supervisions for Multi-Label Image Classification (CVPR 2017 · similarity 0.90)
- Multi-Label Learning From Single Positive Labels (CVPR 2021 · similarity 0.89)
- Deep Semantic Dictionary Learning for Multi-label Image Classification (AAAI 2021 · similarity 0.89)
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.