Papers › Learning Semantic Concepts and Order for Image and Sentence Matching

Learning Semantic Concepts and Order for Image and Sentence Matching

6 Dec 2017CVPR 2018 6arXiv:1712.02036archive 2025-07-28

Yan Huang, Qi Wu, Liang Wang

Image and sentence matching has made great progress recently, but it remains challenging due to the large visual-semantic discrepancy. This mainly arises from that the representation of pixel-level image usually lacks of high-level semantic information as in its matched sentence. In this work, we propose a semantic-enhanced image and sentence matching model, which can improve the image representation by learning semantic concepts and then organizing them in a correct semantic order. Given an image, we first use a multi-regional multi-label CNN to predict its semantic concepts, including objects, properties, actions, etc. Then, considering that different orders of semantic concepts lead to diverse semantic meanings, we use a context-gated sentence generation scheme for semantic order learning. It simultaneously uses the image global context containing concept relations as reference and the groundtruth semantic order in the matched sentence as supervision. After obtaining the improved image representation, we learn the sentence representation with a conventional LSTM, and then jointly perform image and sentence matching and sentence generation for model learning. Extensive experiments demonstrate the effectiveness of our learned semantic concepts and order, by achieving the state-of-the-art results on two public benchmark datasets.

PaperPDFConference PDF

In Syntology View this paper on Syntology: its repositories, every harvested function with whether it ran, its licence and the call to fetch it.

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

Code

No code repository is listed for this paper in the archive or in Syntology's graph.

Code Syntology ran Syntology

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

Tasks

Cross-Modal RetrievalSentence

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Cross-Modal Retrieval COCO 2014 SCO (ResNet) Image-to-text R@1 42.8 #33 of 36 Archive leaderboard report
Cross-Modal Retrieval COCO 2014 SCO (ResNet) Image-to-text R@10 83.0 #33 of 36 Archive leaderboard report
Cross-Modal Retrieval COCO 2014 SCO (ResNet) Image-to-text R@5 72.3 #33 of 36 Archive leaderboard report
Cross-Modal Retrieval COCO 2014 SCO (ResNet) Text-to-image R@1 33.1 #33 of 36 Archive leaderboard report
Cross-Modal Retrieval COCO 2014 SCO (ResNet) Text-to-image R@10 75.5 #33 of 36 Archive leaderboard report
Cross-Modal Retrieval COCO 2014 SCO (ResNet) Text-to-image R@5 62.9 #33 of 36 Archive leaderboard report
Cross-Modal Retrieval Flickr30k SCO (ResNet) Image-to-text R@1 55.5 #23 of 27 Archive leaderboard report
Cross-Modal Retrieval Flickr30k SCO (ResNet) Image-to-text R@10 89.3 #23 of 27 Archive leaderboard report
Cross-Modal Retrieval Flickr30k SCO (ResNet) Image-to-text R@5 82.0 #23 of 27 Archive leaderboard report
Cross-Modal Retrieval Flickr30k SCO (ResNet) Text-to-image R@1 41.1 #23 of 27 Archive leaderboard report
Cross-Modal Retrieval Flickr30k SCO (ResNet) Text-to-image R@10 80.1 #23 of 27 Archive leaderboard report
Cross-Modal Retrieval Flickr30k SCO (ResNet) Text-to-image R@5 70.5 #23 of 27 Archive leaderboard report
Image Retrieval Flickr30K 1K test SCO R@1 41.1 #11 of 18 Archive leaderboard report
Image Retrieval Flickr30K 1K test SCO R@10 80.1 #11 of 18 Archive leaderboard report
Image Retrieval Flickr30K 1K test SCO R@5 70.5 #11 of 18 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

1x1 ConvolutionAverage PoolingBatch NormalizationBottleneck Residual BlockConvolutionGlobal Average PoolingKaiming InitializationLSTMMax PoolingReLUResidual BlockResidual ConnectionSigmoid ActivationTanh Activation

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