Search

Heavy computation is a bottleneck limiting deep-learningbased feature matching algorithms to be applied in many realtime 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., Comment: Have been accepted by AAAI 2023

Along with the advancement of the process of GNSS modernization, BOC modulation has been widely used. Tough it has a certain anti-multipath capability, reflected BOC signals with the large amplitude and short delay cannot be suppressed. Additionally, some of the traditional multipath suppression algorithms are no longer applicable for the BOC signals, so the multipath mitigation algorithms for BOC signals need to be studied further. CCRW is a kind of anti-multipath algorithm for BOC modulation, but it unsuitable for BOC signals with odd coefficient. In order to overcome its shortcoming, this paper proposes an improved CCRW algorithm for BOC signals with odd modulation coefficient, and simulation results prove the effectiveness of the proposed algorithm.