Human action recognition using distance transform and entropy based features
- PDF / 7,098,521 Bytes
- 27 Pages / 439.642 x 666.49 pts Page_size
- 90 Downloads / 244 Views
Human action recognition using distance transform and entropy based features P. Ramya1 · R. Rajeswari1 Received: 19 October 2019 / Revised: 6 October 2020 / Accepted: 23 October 2020 / © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Human action recognition based on silhouette images has wide applications in computer vision, human computer interaction and intelligent surveillance. It is a challenging task due to the complex actions in nature. In this paper, a human action recognition method is proposed which is based on the distance transform and entropy features of human silhouettes. In the first stage, background subtraction is performed by applying correlation coefficient based frame difference technique to extract silhouette images. In the second stage, distance transform based features and entropy features are extracted from the silhouette images. The distance transform based features and entropy features provide the shape and local variation information. These features are given as input to neural networks to recognize various human actions. The proposed method is tested on three different datasets viz., Weizmann, KTH and UCF50. The proposed method obtains an accuracy of 92.5%, 91.4% and 80% for Weizmann, KTH and UCF50 datasets respectively. The experimental results show that the proposed method for human action recognition is comparable to other state-of-the-art human action recognition methods. Keywords Human action recognition · Silhouettes · Distance transform · Entropy · Neural networks
1 Introduction Over the last two decades, learning, understanding and distinguishing various types of human actions from video sequences have become major research problems in computer vision [66]. Appetite on human action recognition methods are developing very fast and these methods are used in many real-time applications, such as video surveillance [21, 51], video retrieval/indexing [17, 65], content based summarization [67], assisted living [38], analysis of sports videos [31], health monitoring [24] and, analysis of shopping behaviour [3]. R. Rajeswari
[email protected] 1
Department of Computer Applications, Bharathiar University, Coimbatore, Tamilnadu, India
Multimedia Tools and Applications
The aim of human action recognition is to classify the unlabelled video frames into simple and/ or complex human actions such as walking, boxing, kicking, punching and hand waving. Several survey papers are published in the literature [10, 15, 20, 35, 43], each on underlining a particular characteristic of recognizing human actions. Although a lot of global and local feature based methods have been proposed for recognizing human actions, it is still a challenging task. Particularly, perfect silhouette feature extraction in complex scenes has recently attracted much attention [8, 25] as they are able to effectively represent shape information. Hence, in this work silhouette based features are used to perform human action recognition. Moreover, using silhouette based features helps in elim
Data Loading...
