8 March 2017 in Scientific Reports
We establish two complementarity relations for the relative entropy of coherence in quantum information processing, i.e., quantum dense coding and teleportation. We first give an uncertainty-like expression relating local quantum coherence to the capacity of optimal dense coding for bipartite system. The relation can also be applied to the case of dense coding by using unital memoryless noisy quantum channels. Further, the relation between local quantum coherence and teleportation fidelity for two-qubit system is given.
1 June 2017 in Physica B: Condensed Matter
4 August 2022 in IET Computer Vision
Few‐shot action recognition aims to learn a classification model with good generalisation ability when trained with only a few labelled videos. However, it is difficult to learn discriminative feature representations for videos in such a setting. The Elastic Temporal Alignment (ETA) for few‐shot action recognition is proposed. First, a convolutional neural network is employed to extract feature representations of video frames sparsely sampled from videos. In order to obtain the similarity of two videos, a temporal alignment estimation function is utilised to estimate the matching score between each pair of frames from the two videos through an elastic alignment mechanism. The analysis shows that when we judge whether two frames from respective videos are matched, multiple adjacent frames in the videos should be considered, so as to embody the temporal information. Thus, before feeding per‐frame feature vectors of videos into the temporal alignment estimation function, a temporal message passing function is leveraged to propagate the information of per‐frame features in the temporal domain. The method has been evaluated on four action recognition datasets, including Kinetics, Something‐Something V2, HMDB51, and UCF101. The experimental results verify the effectiveness of ETA and show its superiority over state‐of‐the‐art methods.
1 May 2021 in IOP Conference Series: Earth and Environmental Science
In view of the difficulty of sensor installation and complicated process in traditional structure health monitoring, a feature point monitoring based on optical flow is proposed in this paper. A high-speed camera is applied to capture the dynamic response of the structure under excitation, and feature point could be selected and marked at any position in the video. The movement of the selected point could be traced and calculated by feature optical flow method, and the displacement time series of the feature point could be obtained. Results show that the displacement time series obtained from proposed method in this paper is consistent with the data of the pull-wire displacement sensor, the maximum displacement error is 3.3%, the method proposed is more flexible, convenient and of low cost.