Papers › GRIT: Faster and Better Image captioning Transformer Using Dual Visual Features

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

20 Jul 2022arXiv:2207.09666archive 2025-07-28

Van-Quang Nguyen, Masanori Suganuma, Takayuki Okatani

Current state-of-the-art methods for image captioning employ region-based features, as they provide object-level information that is essential to describe the content of images; they are usually extracted by an object detector such as Faster R-CNN. However, they have several issues, such as lack of contextual information, the risk of inaccurate detection, and the high computational cost. The first two could be resolved by additionally using grid-based features. However, how to extract and fuse these two types of features is uncharted. This paper proposes a Transformer-only neural architecture, dubbed GRIT (Grid- and Region-based Image captioning Transformer), that effectively utilizes the two visual features to generate better captions. GRIT replaces the CNN-based detector employed in previous methods with a DETR-based one, making it computationally faster. Moreover, its monolithic design consisting only of Transformers enables end-to-end training of the model. This innovative design and the integration of the dual visual features bring about significant performance improvement. The experimental results on several image captioning benchmarks show that GRIT outperforms previous methods in inference accuracy and speed.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

davidnvq/grit officialmentioned in papermentioned on GitHubpytorch report
pacman-ctm/thesis_code mentioned on GitHub report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Image Captioning

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Captioning COCO Captions GRIT (No VL pretraining - base) BLEU-1 84.2 #9 of 41 Archive leaderboard report
Image Captioning COCO Captions GRIT (No VL pretraining - base) BLEU-4 42.4 #9 of 41 Archive leaderboard report
Image Captioning COCO Captions GRIT (No VL pretraining - base) CIDER 144.2 #9 of 41 Archive leaderboard report
Image Captioning COCO Captions GRIT (No VL pretraining - base) METEOR 30.6 #9 of 41 Archive leaderboard report
Image Captioning COCO Captions GRIT (No VL pretraining - base) ROUGE-L 60.7 #9 of 41 Archive leaderboard report
Image Captioning COCO Captions GRIT (No VL pretraining - base) SPICE 24.3 #9 of 41 Archive leaderboard report
Image Captioning nocaps in-domain GRIT (zero-shot, no VL pretraining, no CBS) CIDEr 105.9 #8 of 41 Archive leaderboard report
Image Captioning nocaps in-domain GRIT (zero-shot, no VL pretraining, no CBS) SPICE 13.6 #8 of 41 Archive leaderboard report
Image Captioning nocaps out-of-domain GRIT (zero-shot, no CBS, no VL pretraining, single model) CIDEr 72.6 #19 of 40 Archive leaderboard report
Image Captioning nocaps out-of-domain GRIT (zero-shot, no CBS, no VL pretraining, single model) SPICE 11.1 #19 of 40 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

ConvolutionFaster R-CNNRPNRoIPoolSoftmax

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections