TY - JOUR
T1 - Histogram of Fuzzy Local Spatio-Temporal Descriptors for Video Action Recognition
AU - Zuo, Zheming
AU - Yang, Longzhi
AU - Liu, Yonghuai
AU - Chao, Fei
AU - Song, Ran
AU - Qu, Yanpeng
PY - 2020/6/1
Y1 - 2020/6/1
N2 - Feature extraction plays a vital role in visual action recognition. Many existing gradient-based feature extractors, including histogram of oriented gradients, histogram of optical flow, motion boundary histograms, and histogram of motion gradients, build histograms for representing different actions over the spatiooral 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 article proposes a novel method to handle such uncertainty by fuzzifying these feature extractors. The proposed approach has two advantages: it better represents the ambiguous boundaries between the bins and, thus, the fuzziness of the spatiooral visual information entailed in videos; and 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 are evaluated on two publicly available datasets, demonstrating that the proposed approach outperforms the original counterparts and other state-of-the-art methods.
AB - Feature extraction plays a vital role in visual action recognition. Many existing gradient-based feature extractors, including histogram of oriented gradients, histogram of optical flow, motion boundary histograms, and histogram of motion gradients, build histograms for representing different actions over the spatiooral 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 article proposes a novel method to handle such uncertainty by fuzzifying these feature extractors. The proposed approach has two advantages: it better represents the ambiguous boundaries between the bins and, thus, the fuzziness of the spatiooral visual information entailed in videos; and 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 are evaluated on two publicly available datasets, demonstrating that the proposed approach outperforms the original counterparts and other state-of-the-art methods.
KW - Action recognition
KW - fuzziness
KW - histogram
KW - local feature descriptors
KW - video feature extraction
U2 - 10.1109/TII.2019.2957268
DO - 10.1109/TII.2019.2957268
M3 - Article
AN - SCOPUS:85081995271
VL - 16
SP - 4059
EP - 4067
JO - IEEE Transactions on Industrial Informatics
JF - IEEE Transactions on Industrial Informatics
SN - 1551-3203
IS - 6
M1 - 8919994
ER -