Papers › A Deep Local and Global Scene-Graph Matching for Image-Text Retrieval

A Deep Local and Global Scene-Graph Matching for Image-Text Retrieval

4 Jun 2021arXiv:2106.02400archive 2025-07-28

Manh-Duy Nguyen, Binh T. Nguyen, Cathal Gurrin

Conventional approaches to image-text retrieval mainly focus on indexing visual objects appearing in pictures but ignore the interactions between these objects. Such objects occurrences and interactions are equivalently useful and important in this field as they are usually mentioned in the text. Scene graph presentation is a suitable method for the image-text matching challenge and obtained good results due to its ability to capture the inter-relationship information. Both images and text are represented in scene graph levels and formulate the retrieval challenge as a scene graph matching challenge. In this paper, we introduce the Local and Global Scene Graph Matching (LGSGM) model that enhances the state-of-the-art method by integrating an extra graph convolution network to capture the general information of a graph. Specifically, for a pair of scene graphs of an image and its caption, two separate models are used to learn the features of each graph's nodes and edges. Then a Siamese-structure graph convolution model is employed to embed graphs into vector forms. We finally combine the graph-level and the vector-level to calculate the similarity of this image-text pair. The empirical experiments show that our enhancement with the combination of levels can improve the performance of the baseline method by increasing the recall by more than 10% on the Flickr30k dataset.

PaperPDFCode

Code

m2man/LGSGM 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

Graph MatchingImage RetrievalImage-text RetrievalImage-text matchingImage-to-Text RetrievalRetrievalText MatchingText Retrieval

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Retrieval Flickr30K 1K test LGSGM R@1 57.4 #4 of 18 Archive leaderboard report
Image Retrieval Flickr30K 1K test LGSGM R@10 90.2 #4 of 18 Archive leaderboard report
Image Retrieval Flickr30K 1K test LGSGM R@5 84.1 #4 of 18 Archive leaderboard report
Image Retrieval Flickr30k LGSGM Recall@1 57.4 #7 of 9 Archive leaderboard report
Image Retrieval Flickr30k LGSGM Recall@10 90.2 #7 of 9 Archive leaderboard report
Image Retrieval Flickr30k LGSGM Recall@5 84.1 #7 of 9 Archive leaderboard report
Image Retrieval Flickr30k LGSGM Recall@Sum 231.7 #7 of 9 Archive leaderboard report
Image Retrieval Flickr30k GSMN Recall@10 89 #8 of 9 Archive leaderboard report
Image Retrieval Flickr30k GSMN Recall@5 82.3 #8 of 9 Archive leaderboard report
Image Retrieval Flickr30k GSMN Recall@Sum 228.7 #8 of 9 Archive leaderboard report
Image-to-Text Retrieval Flickr30k GSMN Recall@1 76.4 #10 of 11 Archive leaderboard report
Image-to-Text Retrieval Flickr30k GSMN Recall@10 97.3 #10 of 11 Archive leaderboard report
Image-to-Text Retrieval Flickr30k GSMN Recall@5 94.3 #10 of 11 Archive leaderboard report
Image-to-Text Retrieval Flickr30k GSMN Recall@Sum 268 #10 of 11 Archive leaderboard report
Image-to-Text Retrieval Flickr30k LGSGM Recall@1 71 #11 of 11 Archive leaderboard report
Image-to-Text Retrieval Flickr30k LGSGM Recall@10 96.1 #11 of 11 Archive leaderboard report
Image-to-Text Retrieval Flickr30k LGSGM Recall@5 91.9 #11 of 11 Archive leaderboard report
Image-to-Text Retrieval Flickr30k LGSGM Recall@Sum 259 #11 of 11 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

Convolution

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