Research map / Vision-language and multimodal / Human and embodied modeling
Reflective Decoding Network for Image Captioning
Lei Ke, Wenjie Pei, Ruiyu Li, Xiaoyong Shen, Yu-Wing Tai
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 reflective position module to the baseline decoder (ablation) | RDN baseline · COCO Karpathy test | CIDEr: 113.2 -> 114.3 | Improved |
| add reflective attention module to the baseline decoder (ablation) | RDN baseline · COCO Karpathy test | CIDEr: 113.2 -> 115.0 | Improved |
| combine reflective attention and position modules (ablation) | RDN baseline · COCO Karpathy test | CIDEr: 113.2 -> 115.3 | Improved |
| combine both modules (ablation) | RDN baseline · COCO Karpathy test | BLEU-4: 36.1 -> 36.8 | Improved |
| add position module on top of reflective attention (ablation) | RDN_ref · COCO Karpathy test | CIDEr: 115.0 -> 115.3 | Improved |
| add reflective position module only (ablation) | RDN baseline · COCO Karpathy test SPICE breakdown | SPICE count F-score: 7.4 -> 6.0 | Hurt |
| add both modules (ablation) | RDN baseline · COCO Karpathy test SPICE breakdown | SPICE count F-score: 7.4 -> 8.3 | Improved |
| add reflective attention only (ablation) | RDN baseline · COCO Karpathy test SPICE breakdown | SPICE color F-score: 12.3 -> 11.0 | Hurt |
| RDN decoder versus Up-Down | RDN · COCO Karpathy test (single model) | CIDEr: 113.5 -> 115.3 | Improved |
| RDN versus RFNet on online server (c5) | RDN · COCO test server c5 | CIDEr: 122.9 -> 121.2 | Hurt |
Closest papers by meaning
- An Empirical Study of Language CNN for Image Captioning (ICCV 2017 · similarity 0.89)
- From Show to Tell: A Survey on Deep Learning-Based Image Captioning (TPAMI 2022 · similarity 0.89)
- With a Little Help from Your Own Past: Prototypical Memory Networks for Image Captioning (ICCV 2023 · similarity 0.89)
- Injecting Semantic Concepts Into End-to-End Image Captioning (CVPR 2022 · similarity 0.88)
- Painting with Words: Elevating Detailed Image Captioning with Benchmark and Alignment Learning (ICLR 2025 · similarity 0.88)
- Noise-Aware Image Captioning with Progressively Exploring Mismatched Words (AAAI 2024 · similarity 0.87)
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.
