Stacked Cross Attention for Image-Text Matching

In this paper, we study the problem of image-text matching. Inferring the latent semantic alignment between objects or other salient stuff (e.g. snow, sky, lawn) and the corresponding words in sentences allows to capture fine-grained interplay between vision and language, and makes image-text matching more interpretable. Prior work either simply aggregates the similarity of all possible pairs of regions and words without attending differentially to more and less important words or regions, or uses a multi-step attentional process to capture limited number of semantic alignments which is less interpretable. In this paper, we present Stacked Cross Attention to discover the full latent alignments using both image regions and words in a sentence as context and infer image-text similarity. Our approach achieves the state-of-the-art results on the MS-COCO and Flickr30K datasets. On Flickr30K, our approach outperforms the current best methods by 22.1% relatively in text retrieval from image query, and 18.2% relatively in image retrieval with text query (based on Recall@1). On MS-COCO, our approach improves sentence retrieval by 17.8% relatively and image retrieval by 16.6% relatively (based on Recall@1 using the 5K test set). Code has been made available at: https://github.com/kuanghuei/SCAN.

PDF Abstract ECCV 2018 PDF ECCV 2018 Abstract

Datasets


Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Cross-Modal Retrieval COCO 2014 SCAN Image-to-text R@1 50.4 # 13
Image-to-text R@10 90.0 # 11
Image-to-text R@5 82.2 # 12
Text-to-image R@1 38.6 # 14
Text-to-image R@10 80.4 # 12
Text-to-image R@5 69.3 # 13
Cross-Modal Retrieval Flickr30k SCAN Image-to-text R@1 67.4 # 9
Image-to-text R@10 95.8 # 9
Image-to-text R@5 90.3 # 9
Text-to-image R@1 48.6 # 9
Text-to-image R@10 85.2 # 9
Text-to-image R@5 77.7 # 9
Image Retrieval Flickr30K 1K test SCAN i-t R@1 44.0 # 9
R@10 82.6 # 9
R@5 74.2 # 9

Methods


No methods listed for this paper. Add relevant methods here