Research map / Vision-language and multimodal / Human and embodied modeling
Recurrent Fusion Network for Image captioning
Wenhao Jiang, Lin Ma, Yu-Gang Jiang, Wei Liu, Tong Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add fusion stage I (ablation) | RFNet · MSCOCO Karpathy test (cross-entropy, beam 3) | CIDEr: 108.9 -> 112.5 | Improved |
| add fusion stage II (ablation) | RFNet · MSCOCO Karpathy test (cross-entropy, beam 3) | CIDEr: 111.5 -> 112.5 | Improved |
| add interactions among stage I components (ablation) | RFNet · MSCOCO Karpathy test (cross-entropy, beam 3) | CIDEr: 112.1 -> 112.5 | Improved |
| add discriminative supervision lambda=1 (ablation) | RFNet · MSCOCO test (greedy, lambda sweep) | CIDEr: 105.2 -> 107.2 | Improved |
| add discriminative supervision lambda=10 (ablation) | RFNet · MSCOCO test (greedy, lambda sweep) | CIDEr: 105.2 -> 107.3 | Improved |
| raise discriminative supervision weight to 100 (ablation) | RFNet · MSCOCO test (greedy, lambda sweep) | CIDEr: 107.3 -> 104.7 | Hurt |
| add ResNet (F1) encoder to the other four (ablation) | RFNet · MSCOCO val (greedy, encoder subsets) | CIDEr: 104.8 -> 107.15 | Improved |
| add DenseNet (F4) encoder to the other four (ablation) | RFNet · MSCOCO val (greedy, encoder subsets) | CIDEr: 107.10 -> 107.15 | No clear effect |
| add Inception-V3 (F3) encoder to the other four (ablation) | RFNet · MSCOCO val (greedy, encoder subsets) | CIDEr: 106.90 -> 107.15 | No clear effect |
| add Inception-ResNet-V2 (F5) encoder to the other four (ablation) | RFNet · MSCOCO val (greedy, encoder subsets) | CIDEr: 106.72 -> 107.15 | Improved |
| fuse five encoders rather than two (ablation) | RFNet · MSCOCO val (greedy, encoder subsets) | CIDEr: 106.35 -> 107.15 | Improved |
| ensemble of 4 RFNets instead of single model | RFNet · MSCOCO Karpathy test (cross-entropy, beam 3) | CIDEr: 112.5 -> 116.3 | Improved |
Closest papers by meaning
- An Empirical Study of Language CNN for Image Captioning (ICCV 2017 · similarity 0.90)
- Entangled Transformer for Image Captioning (ICCV 2019 · similarity 0.90)
- Convolutional Image Captioning (CVPR 2018 · similarity 0.90)
- Temporal Deformable Convolutional Encoder-Decoder Networks for Video Captioning (AAAI 2019 · similarity 0.89)
- Reflective Decoding Network for Image Captioning (ICCV 2019 · similarity 0.88)
- Incorporating Copying Mechanism in Image Captioning for Learning Novel Objects (CVPR 2017 · similarity 0.88)
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.
