Papers › Beyond a Pre-Trained Object Detector: Cross-Modal Textual and Visual Context for Image...

Beyond a Pre-Trained Object Detector: Cross-Modal Textual and Visual Context for Image Captioning

9 May 2022CVPR 2022 1arXiv:2205.04363archive 2025-07-28

Chia-Wen Kuo, Zsolt Kira

Significant progress has been made on visual captioning, largely relying on pre-trained features and later fixed object detectors that serve as rich inputs to auto-regressive models. A key limitation of such methods, however, is that the output of the model is conditioned only on the object detector's outputs. The assumption that such outputs can represent all necessary information is unrealistic, especially when the detector is transferred across datasets. In this work, we reason about the graphical model induced by this assumption, and propose to add an auxiliary input to represent missing information such as object relationships. We specifically propose to mine attributes and relationships from the Visual Genome dataset and condition the captioning model on them. Crucially, we propose (and show to be important) the use of a multi-modal pre-trained model (CLIP) to retrieve such contextual descriptions. Further, object detector models are frozen and do not have sufficient richness to allow the captioning model to properly ground them. As a result, we propose to condition both the detector and description outputs on the image, and show qualitatively and quantitatively that this can improve grounding. We validate our method on image captioning, perform thorough analyses of each component and importance of the pre-trained multi-modal model, and demonstrate significant improvements over the current state of the art, specifically +7.5% in CIDEr and +1.3% in BLEU-4 metrics.

PaperPDFConference PDFCode

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

Code

GT-RIPL/Xmodal-Ctx officialmentioned on GitHubpytorch 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 CaptioningObject

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Captioning COCO Captions Xmodal-Ctx BLEU-1 83.4 #12 of 41 Archive leaderboard report
Image Captioning COCO Captions Xmodal-Ctx BLEU-4 41.4 #12 of 41 Archive leaderboard report
Image Captioning COCO Captions Xmodal-Ctx CIDER 139.9 #12 of 41 Archive leaderboard report
Image Captioning COCO Captions Xmodal-Ctx METEOR 30.4 #12 of 41 Archive leaderboard report
Image Captioning COCO Captions Xmodal-Ctx ROUGE-L 60.4 #12 of 41 Archive leaderboard report
Image Captioning COCO Captions Xmodal-Ctx SPICE 24.0 #12 of 41 Archive leaderboard report
Image Captioning COCO Captions Xmodal-Ctx + OSCAR BLEU-4 41.3 #13 of 41 Archive leaderboard report
Image Captioning COCO Captions Xmodal-Ctx + OSCAR CIDER 142.2 #13 of 41 Archive leaderboard report
Image Captioning COCO Captions Xmodal-Ctx + OSCAR SPICE 24.9 #13 of 41 Archive leaderboard report
Image Captioning COCO Captions Xmodal-Ctx BLEU-1 81.5 #21 of 41 Archive leaderboard report
Image Captioning COCO Captions Xmodal-Ctx BLEU-4 39.7 #21 of 41 Archive leaderboard report
Image Captioning COCO Captions Xmodal-Ctx CIDER 135.9 #21 of 41 Archive leaderboard report
Image Captioning COCO Captions Xmodal-Ctx METEOR 30.0 #21 of 41 Archive leaderboard report
Image Captioning COCO Captions Xmodal-Ctx ROUGE-L 59.5 #21 of 41 Archive leaderboard report
Image Captioning COCO Captions Xmodal-Ctx SPICE 23.7 #21 of 41 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.

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