Download pdf
Wiki-match: A multi-model pipeline for image-caption matching task on Wikipedia dataset
- 1 School of Computing and Information System, Singapore Management University, 188065, Singapore
- 2 School of Computer Science, Sichuan University, 610207, China
- 3 School of Data Science, The Chinese University of Hongkong (Shenzhen), 518172, China
* Author to whom correspondence should be addressed.
Abstract
We propose a multi-model pipeline for image-caption matching tasks on Wikipedia-based dataset which leverages object-detection technique and attention mechanism to achieve fine-grained matching between textual representation and image representation. Different from the prior research, we not only evaluate our pipeline effectiveness on common benchmark dataset such as MS-COCO and Flickr30k, but also a new dataset that is de-rived from Wikipedia which is rich in natural entities and abstract concepts. Our findings show: 1) our model pipeline improves R@1 by 113.4%, R@3 by 86.1%, and R@5 by 114.4% compared to the original pipeline provided by the Wikipedia-based dataset. 2) our model pipeline has close to the state-of-the-art performance in common benchmark dataset including Flickr30k and MS-COCO. 3) images that are from Wikipedia creates bigger challenges for models to understand compares to MS-COCO or Flickr30k due to the abstract concepts and broad topics covered by Wikipedia.
Keywords
deep learning, natural language processing, computer vision, attention
[1]. Mei, T., Zhang, W., & Yao, T. (2020). Vision and language: from visual perception to content creation. APSIPA Transactions on Signal and Information Processing, 9:1-8.
[2]. Szeliski, R. (2010). A brief history. In: Gries D. Schneider F.B. (Eds.), Computer vision: algorithms and applications. Springer Science & Business Media. New York; London. pp. 11.
[3]. Li, L. H., Yatskar, M., Yin, D., Hsieh, C. J., & Chang, K. W. (2019). Visualbert: A simple and performant baseline for vision and language. arXiv preprint arXiv:1908.03557.
[4]. Ordonez, V., Kulkarni, G., & Berg, T. (2011). Im2text: Describing images using 1 million captioned photographs. Advances in neural information processing systems, 24: 1143-1151.
[5]. Hodosh, M., Young, P., & Hockenmaier, J. (2013). Framing image description as a ranking task: Data, models and evaluation metrics. Journal of Artificial Intelligence Research, 47: 853-899.
[6]. Wikimedia Foundation. (2021). Wikipedia - image/caption matching. https://www.kaggle.com/c/wikipedia-image-caption.
[7]. Ma, L., Lu, Z., Shang, L., & Li, H. (2015). Multimodal convolutional neural networks for matching image and sentence. In: Proceedings of the IEEE international conference on computer vision. pp. 2623-2631.
[8]. Wang, L., Li, Y., Huang, J., & Lazebnik, S. (2018). Learning two-branch neural networks for image-text matching tasks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 41: 394-407.
[9]. Xu, X., Wang, T., Yang, Y., Zuo, L., Shen, F., & Shen, H. T. (2020). Cross-modal attention with semantic consistence for image–text matching. IEEE transactions on neural networks and learning systems, 31: 5412-5425.
[10]. Wang, D., Wang, L., Song, S., Huang, G., Guo, Y., Cheng, S., ... & Du, A. (2021). Fusion layer attention for image-text matching. Neurocomputing, 442: 249-259.
[11]. Nam, H., Ha, J. W., & Kim, J. (2017). Dual attention networks for multimodal reasoning and matching. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 299-307.
[12]. Lee, K. H., Chen, X., Hua, G., Hu, H., & He, X. (2018). Stacked cross attention for image-text matching. In: Proceedings of the European Conference on Computer Vision (ECCV). pp. 201-216.
[13]. Huang, F., Zhang, X., Zhao, Z., & Li, Z. (2018). Bi-directional spatial-semantic attention networks for image-text matching. IEEE Transactions on Image Processing, 28: 2008-2020.
[14]. Girshick, R. (2015). Fast r-cnn. In: Proceedings of the IEEE international conference on computer vision, pp. 1440-1448.
[15]. Ren, S., He, K., Girshick, R., & Sun, J. (2015). Faster R-CNN: Towards real-time object detection with region proposal networks. Advances in neural information processing systems, 28: 91-99.
[16]. Wang, S., Zhou, W., & Jiang, C. (2020). A survey of word embeddings based on deep learning. Computing, 102: 717-740.
[17]. Peters, M. E., Neumann, M., Iyyer, M., Gardner, M., Clark, C., Lee, K., & Zettlemoyer, L. (2018). Deep contextualized word representations. arXiv 2018. arXiv preprint arXiv:1802.05365, 12.
[18]. Srinivasan, K., Raman, K., Chen, J., Bendersky, M., & Najork, M. (2021). WIT: Wikipedia-based image text dataset for multimodal multilingual machine learning. arXiv preprint arXiv:2103.01913.
[19]. Faghri, F., Fleet, D. J., Kiros, J. R., & Fidler, S. (2017). Vse++: Improving visual-semantic embeddings with hard negatives. arXiv preprint arXiv:1707.05612.
[20]. Wang, Y., Yang, H., Qian, X., Ma, L., Lu, J., Li, B., & Fan, X. (2019). Position focused attention network for image-text matching. arXiv preprint arXiv:1907.09748.
Cite this article
Chen,Y.;Lei,S.;Sun,Z. (2023). Wiki-match: A multi-model pipeline for image-caption matching task on Wikipedia dataset. Applied and Computational Engineering,2,45-55.
Data availability
The datasets used and/or analyzed during the current study will be available from the authors upon reasonable request.
Disclaimer/Publisher's Note
The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of EWA Publishing and/or the editor(s). EWA Publishing and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
About volume
Volume title: Proceedings of the 4th International Conference on Computing and Data Science (CONF-CDS 2022)
© 2024 by the author(s). Licensee EWA Publishing, Oxford, UK. This article is an open access article distributed under the terms and
conditions of the Creative Commons Attribution (CC BY) license. Authors who
publish this series agree to the following terms:
1. Authors retain copyright and grant the series right of first publication with the work simultaneously licensed under a Creative Commons
Attribution License that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this
series.
2. Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the series's published
version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial
publication in this series.
3. Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and
during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See
Open access policy for details).