Research map / Vision-language and multimodal / Human and embodied modeling
An Empirical Study of Language CNN for Image Captioning
Jiuxiang Gu, Gang Wang, Jianfei Cai, Tsuhan Chen
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 language CNN to simple RNN | Simple RNN · MS COCO | CIDEr: 87.0 -> 95.2 | Improved |
| add language CNN to RHN | RHN · MS COCO | CIDEr: 90.6 -> 98.9 | Improved |
| add language CNN to LSTM | LSTM · MS COCO | CIDEr: 92.6 -> 99.1 | Improved |
| add language CNN to GRU | GRU · MS COCO | CIDEr: 93.3 -> 96.1 | Improved |
| add language CNN to LSTM | LSTM · Flickr30k | CIDEr: 35.0 -> 45.0 | Improved |
| add language CNN to RHN | RHN · Flickr30k | BLEU-4: 20.0 -> 30.7 | Improved |
| add language CNN to simple RNN | Simple RNN · Flickr30k | CIDEr: 32.5 -> 65.4 | Improved |
| add a 2nd LSTM layer instead of a language CNN (ablation) | LSTM · MS COCO | CIDEr: 92.6 -> 93.2 | Improved |
| add a 3rd LSTM layer (ablation) | LSTM · MS COCO | CIDEr: 92.6 -> 92.9 | No clear effect |
| add RNN on top of the language CNN (ablation) | CNN_L · MS COCO | CIDEr: 56.8 -> 95.2 | Improved |
| use a larger window of previous words (2 -> 8 words) (ablation) | CNN_L+RHN · MS COCO | CIDEr: 93.8 -> 95.9 | Improved |
| use full 16-word language CNN rather than 8 words (ablation) | CNN_L+RHN · MS COCO | CIDEr: 95.9 -> 98.9 | Improved |
Closest papers by meaning
- Convolutional Image Captioning (CVPR 2018 · similarity 0.92)
- From Show to Tell: A Survey on Deep Learning-Based Image Captioning (TPAMI 2022 · similarity 0.91)
- Show and Tell: Lessons Learned from the 2015 MSCOCO Image Captioning Challenge (TPAMI 2016 · similarity 0.91)
- Recurrent Fusion Network for Image captioning (ECCV 2018 · similarity 0.90)
- Image Captioning With Semantic Attention (CVPR 2016 · similarity 0.89)
- Unsupervised Image Captioning (CVPR 2019 · 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.
