Skeleton-based action recognition has attracted the attention of many researchers. In the process of extracting skeleton features, most methods use the first-order features (the joint position of the skeleton) or the second-order features (the length and direction of the skeleton) to represent the skeleton, while ignoring the importance of skeleton semantic features to action. In this paper, a new skeleton appearance semantic feature is proposed to describe the appearance feature of human skeleton. An attention mechanism of skeleton appearance semantic features is further proposed, which integrates channel features to highlight the overall difference. In addition, a method is used to integrate the semantic features of skeleton appearance with the position and velocity features of skeleton joints, as well as the constructed joints type and frame index, so as to improve the representation method of skeleton semantic features. The experiments were carried out in two public skeleton action recognition data sets (NTU-RGB+D 60 and NTU-RGB+D 120), and the recognition accuracy was higher than that of the baseline model SGN.
The end-to-end model for automatic speech recognition (ASR) has recently gained significant interest in the research community. Standard RNN-T loss causes an unreasonable alignment of output and labels, such as the model outputting a sequence of labels at a single time frame and a single label can only be emitted by a single frame. We apply the CTC loss function to RNN-T model to address these problems. On the assumption that the output sequence is no longer than the input sequence we propose an improved forward-backward algorithm to calculate the loss. Experimental results for speech recognition are provided on the AISHELL-1 speech corpus. The results show that TCL converges faster without causing gradient explosion compared with RNN-T.TCL achieves a 9.5% character error rate on a test set of AISHELL-1 Chinese speech corpus, outperforming both the baseline RNN-T reported by Z. Tian etc. and standard CTC achieve11.82% and 10.8% respectively. RNN-T with CTC loss achieves a 19.63% relative reduction in character error rate compared with the standard RNN-T and a 12.04% relative reduction in character error rate compared with standard CTC on the test set.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.