Histogram of Fuzzy Local Spatio-Temporal Descriptors for Video Action Recognition

Zuo, Zheming, Yang, Longzhi, Yonghuai, Liu, Chao, Fei, Song, Ran and Qu, Yanpeng (2020) Histogram of Fuzzy Local Spatio-Temporal Descriptors for Video Action Recognition. IEEE Transactions on Industrial Informatics, 16 (6). pp. 4059-4067. ISSN 1551-3203

FINAL VERSION_IEEE Transactions.pdf - Accepted Version

Download (8MB) | Preview
Official URL: https://doi.org/10.1109/TII.2019.2957268


Feature extraction plays a vital role in visual action recognition. Many existing gradient-based feature extractors, including histogram of oriented gradients (HOG), histogram of optical flow (HOF), motion boundary histograms (MBH), and histogram of motion gradients (HMG), build histograms for representing different actions over the spatio-temporal domain in a video. However, these methods require to set the number of bins for information aggregation in advance. Varying numbers of bins usually lead to inherent uncertainty within the process of pixel voting with regard to the bins in the histogram. This paper proposes a novel method to handle such uncertainty by fuzzifying these feature extractors. The proposed approach has two advantages: i) it better represents the ambiguous boundarie between the bins and thus the fuzziness of th spatio-temporal visual information entailed in videos, and ii) the contribution of each pixel is flexibly controlled by a fuzziness parameter for various scenarios. The proposed family of fuzzy descriptors and a combination of them were evaluate on two publicly available datasets, demonstrating that the proposed approach outperforms the original counterparts and other state-of-the-art methods.

Item Type: Article
Uncontrolled Keywords: Action recognition, fuzziness, histogram, local feature descriptors, video feature extraction
Subjects: G400 Computer Science
Department: Faculties > Engineering and Environment > Computer and Information Sciences
Depositing User: Ay Okpokam
Date Deposited: 03 Dec 2019 12:32
Last Modified: 31 Jul 2021 11:35
URI: http://nrl.northumbria.ac.uk/id/eprint/41657

Actions (login required)

View Item View Item


Downloads per month over past year

View more statistics