An in-depth study of sparse codes on abnormality detection

TY - GEN

T1 - An in-depth study of sparse codes on abnormality detection

AU - Ren, Huamin

AU - Pan, Hong

AU - Olsen, Søren Ingvor

AU - Jensen, Morten Bornø

AU - Moeslund, Thomas B.

N1 - Conference code: 13

PY - 2016

Y1 - 2016

N2 - Sparse representation has been applied successfully in abnormal event detection, in which the baseline is to learn a dictionary accompanied by sparse codes. While much emphasis is put on discriminative dictionary construction, there are no comparative studies of sparse codes regarding abnormality detection. We present an in-depth study of two types of sparse codes solutions-greedy algorithms and convex L1-norm solutions-and their impact on abnormality detection performance. We also propose our framework of combining sparse codes with different detection methods. Our comparative experiments are carried out from various angles to better understand the applicability of sparse codes, including computation time, reconstruction error, sparsity, detection accuracy, and their performance combining various detection methods. The experiment results show that combining OMP codes with maximum coordinate detection could achieve state-of-The-Art performance on the UCSD dataset [14].

AB - Sparse representation has been applied successfully in abnormal event detection, in which the baseline is to learn a dictionary accompanied by sparse codes. While much emphasis is put on discriminative dictionary construction, there are no comparative studies of sparse codes regarding abnormality detection. We present an in-depth study of two types of sparse codes solutions-greedy algorithms and convex L1-norm solutions-and their impact on abnormality detection performance. We also propose our framework of combining sparse codes with different detection methods. Our comparative experiments are carried out from various angles to better understand the applicability of sparse codes, including computation time, reconstruction error, sparsity, detection accuracy, and their performance combining various detection methods. The experiment results show that combining OMP codes with maximum coordinate detection could achieve state-of-The-Art performance on the UCSD dataset [14].

UR - http://www.scopus.com/inward/record.url?scp=85003977003&partnerID=8YFLogxK

U2 - 10.1109/AVSS.2016.7738016

DO - 10.1109/AVSS.2016.7738016

M3 - Article in proceedings

AN - SCOPUS:85003977003

SP - 66

EP - 72

BT - 2016 13th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS)

PB - IEEE

T2 - 13th IEEE International Conference on Advanced Video and Signal Based Surveillance

Y2 - 23 August 2016 through 26 August 2016

ER -