ParaFormer: Parallel Attention Transformer for Efficient Feature Matching
Proceedings of the AAAI Conference on Artificial Intelligence · 26.06.2023 · 37 (2) · 1853-1860
DOI: 10.1609/aaai.v37i2.25275 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 26.06.2023. Verfasst von Xiaoyong Lu und 3 weiteren Autor:innen. Die Arbeit wurde bisher 19-mal zitiert (Crossref).
Abstract (Originalsprache)
Heavy computation is a bottleneck limiting deep-learning-based feature matching algorithms to be applied in many real-time applications. However, existing lightweight networks optimized for Euclidean data cannot address classical feature matching tasks, since sparse keypoint based descriptors are expected to be matched. This paper tackles this problem and proposes two concepts: 1) a novel parallel attention model entitled ParaFormer and 2) a graph based U-Net architecture with attentional pooling. First, ParaFormer fuses features and keypoint positions through the concept of amplitude and phase, and integrates self- and cross-attention in a parallel manner which achieves a win-win performance in terms of accuracy and efficiency. Second, with U-Net architecture and proposed attentional pooling, the ParaFormer-U variant significantly reduces computational complexity, and minimize performance loss caused by downsampling. Sufficient experiments on various applications, including homography estimation, pose estimation, and image matching, demonstrate that ParaFormer achieves state-of-the-art performance while maintaining high efficiency. The efficient ParaFormer-U variant achieves comparable performance with less than 50% FLOPs of the existing attention-based models.
Zitieren
APA
Lu, X., Yan, Y., Kang, B., & Du, S. (2023). ParaFormer: Parallel Attention Transformer for Efficient Feature Matching. Proceedings of the AAAI Conference on Artificial Intelligence, 37(2), 1853-1860. https://doi.org/10.1609/aaai.v37i2.25275
Vancouver
Lu X, Yan Y, Kang B, Du S. ParaFormer: Parallel Attention Transformer for Efficient Feature Matching. Proceedings of the AAAI Conference on Artificial Intelligence. 2023;37(2):1853-1860. doi:10.1609/aaai.v37i2.25275
BibTeX
@article{Lu2023,
author = {Lu, Xiaoyong and Yan, Yaping and Kang, Bin and Du, Songlin},
title = {ParaFormer: Parallel Attention Transformer for Efficient Feature Matching},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2023},
volume = {37},
number = {2},
pages = {1853--1860},
doi = {10.1609/aaai.v37i2.25275}
}
EUVIMED