Dynamical hyperparameter optimization via deep reinforcement learning in tracking
dc.contributor.author | Dong, Xingping | |
dc.contributor.author | Shen, Jianbing | |
dc.contributor.author | Wang, Wenguan | |
dc.contributor.author | Shao, Ling | |
dc.contributor.author | Ling, Haibin | |
dc.contributor.author | Porikli, Fatih | |
dc.date.accessioned | 2024-05-13T00:03:38Z | |
dc.date.issued | 2021 | |
dc.date.updated | 2023-01-15T07:16:58Z | |
dc.description.abstract | Hyperparameters are numerical pre-sets whose values are assigned prior to the commencement of a learning process. Selecting appropriate hyperparameters is often critical for achieving satisfactory performance in many vision problems, such as deep learning-based visual object tracking. However, it is often difficult to determine their optimal values, especially if they are specific to each video input. Most hyperparameter optimization algorithms tend to search a generic range and are imposed blindly on all sequences. In this paper, we propose a novel dynamical hyperparameter optimization method that adaptively optimizes hyperparameters for a given sequence using an action-prediction network leveraged on continuous deep Q-learning. Since the observation space for object tracking is significantly more complex than those in traditional control problems, existing continuous deep Q-learning algorithms cannot be directly applied. To overcome this challenge, we introduce an efficient heuristic strategy to handle high dimensional state space, while also accelerating the convergence behavior. The proposed algorithm is applied to improve two representative trackers, a Siamese-based one and a correlation-filter-based one, to evaluate its generalizability. Their superior performances on several popular benchmarks are clearly demonstrated. Our source code is available at https://github.com/shenjianbing/dqltracking. | en_AU |
dc.description.sponsorship | This work was supported in part by the Beijing Natural Science Foundation under Grant 4182056, the CCF-Tencent Open Fund, Zhijiang Lab’s International Talent Fund for Young Professionals, and the Joint Building Program of Beijing Municipal Education Commission. A preliminary version of this work has appeared in CVPR 2018 [1]. | en_AU |
dc.format.mimetype | application/pdf | en_AU |
dc.identifier.issn | 0162-8828 | en_AU |
dc.identifier.uri | http://hdl.handle.net/1885/317450 | |
dc.language.iso | en_AU | en_AU |
dc.publisher | Institute of Electrical and Electronics Engineers (IEEE Inc) | en_AU |
dc.rights | © 2019 The authors | en_AU |
dc.source | IEEE Transactions on Pattern Analysis and Machine Intelligence | en_AU |
dc.subject | Hyperparameters | en_AU |
dc.subject | continuous deep q-learning | en_AU |
dc.subject | reinforcement learning | en_AU |
dc.subject | visual object tracking | en_AU |
dc.title | Dynamical hyperparameter optimization via deep reinforcement learning in tracking | en_AU |
dc.type | Journal article | en_AU |
local.bibliographicCitation.issue | 5 | en_AU |
local.bibliographicCitation.lastpage | 1529 | en_AU |
local.bibliographicCitation.startpage | 1515 | en_AU |
local.contributor.affiliation | Dong, Xingping, Beijing Institute of Technology | en_AU |
local.contributor.affiliation | Shen, Jianbing, Beijing Institute of Technology | en_AU |
local.contributor.affiliation | Wang, Wenguan, Beijing Institute of Technology | en_AU |
local.contributor.affiliation | Shao, Ling, Inception Institute of Artificial Intelligence | en_AU |
local.contributor.affiliation | Ling, Haibin, Stonybrook University | en_AU |
local.contributor.affiliation | Porikli, Fatih, College of Engineering, Computing and Cybernetics, ANU | en_AU |
local.contributor.authoremail | u5405232@anu.edu.au | en_AU |
local.contributor.authoruid | Porikli, Fatih, u5405232 | en_AU |
local.description.embargo | 2099-12-31 | |
local.description.notes | Imported from ARIES | en_AU |
local.identifier.absfor | 400900 - Electronics, sensors and digital hardware | en_AU |
local.identifier.ariespublication | a383154xPUB18968 | en_AU |
local.identifier.citationvolume | 43 | en_AU |
local.identifier.doi | 10.1109/TPAMI.2019.2956703 | en_AU |
local.identifier.scopusID | 2-s2.0-85103804043 | |
local.identifier.thomsonID | WOS:000637533800004 | |
local.identifier.uidSubmittedBy | a383154 | en_AU |
local.publisher.url | https://ieeexplore.ieee.org/ | en_AU |
local.type.status | Published Version | en_AU |
Downloads
Original bundle
1 - 1 of 1
No Thumbnail Available
- Name:
- Dynamical_Hyperparameter_Optimization_via_Deep_Reinforcement_Learning_in_Tracking.pdf
- Size:
- 8.12 MB
- Format:
- Adobe Portable Document Format
- Description: