AI Paper Atlas

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

ECCV 2018 · Vision-language and multimodal › Human and embodied modeling · approach: Neural image captioning · 306 citations (OpenAlex)

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 changedWhereMeasuredEffect
add fusion stage I (ablation)RFNet · MSCOCO Karpathy test (cross-entropy, beam 3)CIDEr: 108.9 -> 112.5Improved
add fusion stage II (ablation)RFNet · MSCOCO Karpathy test (cross-entropy, beam 3)CIDEr: 111.5 -> 112.5Improved
add interactions among stage I components (ablation)RFNet · MSCOCO Karpathy test (cross-entropy, beam 3)CIDEr: 112.1 -> 112.5Improved
add discriminative supervision lambda=1 (ablation)RFNet · MSCOCO test (greedy, lambda sweep)CIDEr: 105.2 -> 107.2Improved
add discriminative supervision lambda=10 (ablation)RFNet · MSCOCO test (greedy, lambda sweep)CIDEr: 105.2 -> 107.3Improved
raise discriminative supervision weight to 100 (ablation)RFNet · MSCOCO test (greedy, lambda sweep)CIDEr: 107.3 -> 104.7Hurt
add ResNet (F1) encoder to the other four (ablation)RFNet · MSCOCO val (greedy, encoder subsets)CIDEr: 104.8 -> 107.15Improved
add DenseNet (F4) encoder to the other four (ablation)RFNet · MSCOCO val (greedy, encoder subsets)CIDEr: 107.10 -> 107.15No clear effect
add Inception-V3 (F3) encoder to the other four (ablation)RFNet · MSCOCO val (greedy, encoder subsets)CIDEr: 106.90 -> 107.15No clear effect
add Inception-ResNet-V2 (F5) encoder to the other four (ablation)RFNet · MSCOCO val (greedy, encoder subsets)CIDEr: 106.72 -> 107.15Improved
fuse five encoders rather than two (ablation)RFNet · MSCOCO val (greedy, encoder subsets)CIDEr: 106.35 -> 107.15Improved
ensemble of 4 RFNets instead of single modelRFNet · MSCOCO Karpathy test (cross-entropy, beam 3)CIDEr: 112.5 -> 116.3Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy