AI Paper Atlas

Research map / Detection and segmentation / CNN features

GRIT: Faster and Better Image Captioning Transformer Using Dual Visual Features

Van-Quang Nguyen, Masanori Suganuma, Takayuki Okatani

ECCV 2022 · Detection and segmentation › CNN features · approach: Local features and context in CNNs · 159 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 end-to-end training of backbone and detector with the caption generator (ablation)GRIT · COCO Karpathy testCIDEr: 139.6 -> 144.2Improved
pretrain detector on four datasets instead of Visual Genome only (ablation)GRIT · COCO Karpathy testCIDEr: 142.3 -> 144.2Improved
add region features from detector pretraining on top of ImageNet grid-only backbone (ablation)GRIT · COCO Karpathy testCIDEr: 135.5 -> 142.3Improved
increase number of region queries from 50 to 150 (ablation)GRIT (VG-trained) · COCO Karpathy testCIDEr: 141.4 -> 142.3Improved
increase number of region queries from 100 to 150 (ablation)GRIT (VG-trained) · COCO Karpathy testBLEU-4: 41.5 -> 41.9No clear effect
add region features to grid-only concatenated cross-attention (ablation)GRIT concatenated attention · COCO Karpathy testCIDEr: 142.1 -> 143.1Improved
use region features alone instead of grid alone (ablation)GRIT concatenated attention · COCO Karpathy testCIDEr: 142.1 -> 142.9Improved
use parallel cross-attention instead of concatenation (ablation)GRIT · COCO Karpathy testCIDEr: 143.1 -> 144.2Improved
use parallel cross-attention instead of sequential (grid then region) (ablation)GRIT · COCO Karpathy testCIDEr: 144.0 -> 144.2Improved
order sequential attention grid first then region rather than region first (ablation)GRIT sequential attention · COCO Karpathy testCIDEr: 143.6 -> 144.0Improved
add sigmoid gating to parallel cross-attention (vs identity) (ablation)GRIT parallel attention · COCO Karpathy testCIDEr: 143.9 -> 144.2Improved
add sigmoid gating to parallel cross-attention (vs identity) (ablation)GRIT parallel attention · COCO Karpathy testBLEU-4: 41.6 -> 42.4Improved

Later papers that compared with this method

4 later accepted papers report +GRIT (20M) as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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