Asymmetric discriminative correlation filters for visual tracking
- PDF / 1,350,049 Bytes
- 18 Pages / 595.276 x 841.89 pts (A4) Page_size
- 97 Downloads / 229 Views
2020 21(10):1467-1484
1467
Frontiers of Information Technology & Electronic Engineering www.jzus.zju.edu.cn; engineering.cae.cn; www.springerlink.com ISSN 2095-9184 (print); ISSN 2095-9230 (online) E-mail: [email protected]
Asymmetric discriminative correlation filters for visual tracking∗ Shui-wang LI, Qian-bo JIANG, Qi-jun ZHAO, Li LU‡ , Zi-liang FENG National Key Laboratory of Fundamental Science on Synthetic Vision, College of Computer Science, Sichuan University, Chengdu 610065, China E-mail: [email protected]; [email protected]; [email protected]; [email protected]; [email protected] Received Sept. 20, 2019; Revision accepted Apr. 12, 2020; Crosschecked Sept. 2, 2020
Abstract: Discriminative correlation filters (DCF) are efficient in visual tracking and have advanced the field significantly. However, the symmetry of correlation (or convolution) operator results in computational problems and does harm to the generalized translation equivariance. The former problem has been approached in many ways, whereas the latter one has not been well recognized. In this paper, we analyze the problems with the symmetry of circular convolution and propose an asymmetric one, which as a generalization of the former has a weak generalized translation equivariance property. With this operator, we propose a tracker called the asymmetric discriminative correlation filter (ADCF), which is more sensitive to translations of targets. Its asymmetry allows the filter and the samples to have different sizes. This flexibility makes the computational complexity of ADCF more controllable in the sense that the number of filter parameters will not grow with the sample size. Moreover, the normal matrix of ADCF is a block matrix with each block being a two-level block Toeplitz matrix. With this well-structured normal matrix, we design an algorithm for multiplying an N × N two-level block Toeplitz matrix by a vector with time complexity O(N logN ) and space complexity O(N ), instead of O(N 2 ). Unlike DCF-based trackers, introducing spatial or temporal regularization does not increase the essential computational complexity of ADCF. Comparative experiments are performed on a synthetic dataset and four benchmarks, including OTB-2013, OTB-2015, VOT-2016, and Temple-Color, and the results show that our method achieves state-of-the-art visual tracking performance. Key words: Visual tracking; Discriminative correlation filter (DCF); Asymmetric DCF (ADCF) https://doi.org/10.1631/FITEE.1900507 CLC number: TP391
1 Introduction Visual tracking is an important and challenging task in the field of computer vision. Discriminative correlation filters (DCFs) have successfully been applied to this problem and rapid advances have been witnessed in recent years (Henriques et al., 2015; Danelljan et al., 2017b; Galoogahi et al., 2017; Sun C ‡ *
Corresponding author
Project supported by the National Natural Science Foundation of China (No. 61773270) and the Key Research and Development Project of Sichuan Province, China (No. 2019YFG0491) ORCID: Shui-wang LI, https://
Data Loading...
