Efficient joint model learning, segmentation and model updating for visual tracking

2022 ◽  
Author(s):  
Wei Han ◽  
Chamara Kasun Liyanaarachchi Lekamalage ◽  
Guang-Bin Huang
Keyword(s):  
Visual Tracking ◽  
Model Updating ◽  
Joint Model ◽  
Model Learning

Self-paced model learning for robust visual tracking

2017 ◽  
Vol 26 (1) ◽  
pp. 013016 ◽  
Author(s):  
Wenhui Huang ◽  
Jason Gu ◽  
Xin Ma ◽  
Yibin Li
Keyword(s):  
Visual Tracking ◽  
Model Learning

Nonlinear joint model updating of strong column-weak beam type frames based on instantaneous characteristics of the responses under earthquake excitations

2017 ◽  
Vol 20 (5) ◽  
pp. 682-693 ◽  
Author(s):  
Ping-Ping Yuan ◽  
Zuo-Cai Wang ◽  
Wei-Xin Ren ◽  
Wen-Yu He
Keyword(s):  
Objective Function ◽  
Simulated Annealing Algorithm ◽  
Model Updating ◽  
Element Model ◽  
Joint Model ◽  
Model Parameters ◽  
Pass Filter ◽  
Low Pass Filter ◽  
Nonlinear Joint ◽  
Instantaneous Frequencies

In this article, a novel method based on the instantaneous frequencies and amplitudes of the principal response components is presented for nonlinear joint model updating. The instantaneous frequencies and amplitudes are first extracted by a low-pass filter with Hilbert transform. Then, limited point values of the extracted instantaneous frequencies and amplitudes are applied to represent the response of the nonlinear structure. Finally, an objective function based on the residuals of instantaneous frequencies and amplitudes between experimental structure and finite element model is established using the response surface method. The optimal values of the nonlinear joint model parameters are obtained by minimizing the objective function using simulated annealing algorithm. To verify the effectiveness of the proposed method, a three-story frame with bilinear moment–rotation relationship at the beam-column joints under earthquake excitations is simulated as a numerical example. The accuracy of the proposed nonlinear joint model updating procedure is quantified using the defined error indices. The effects of the selected data point number and the weight factors in the objective function are also discussed in the article. The results indicate that the proposed method can effectively update the nonlinear joint model with high accuracy even with noise effect.

scholarly journals Multi-Stream Deep Similarity Learning Networks for Visual Tracking

2017 ◽  
Author(s):  
Kunpeng Li ◽  
Yu Kong ◽  
Yun Fu
Keyword(s):  
Visual Tracking ◽  
Model Updating ◽  
Feature Space ◽  
Fine Tuning ◽  
Similarity Learning ◽  
Learning Networks ◽  
Target Template ◽  
Background Clutter ◽  
Cluttered Background ◽  
Background Context

Visual tracking has achieved remarkable success in recent decades, but it remains a challenging problem due to appearance variations over time and complex cluttered background. In this paper, we adopt a tracking-by-verification scheme to overcome these challenges by determining the patch in the subsequent frame that is most similar to the target template and distinctive to the background context. A multi-stream deep similarity learning network is proposed to learn the similarity comparison model. The loss function of our network encourages the distance between a positive patch in the search region and the target template to be smaller than that between positive patch and the background patches. Within the learned feature space, even if the distance between positive patches becomes large caused by the appearance change or interference of background clutter, our method can use the relative distance to distinguish the target robustly. Besides, the learned model is directly used for tracking with no need of model updating, parameter fine-tuning and can run at 45 fps on a single GPU. Our tracker achieves state-of-the-art performance on the visual tracking benchmark compared with other recent real-time-speed trackers, and shows better capability in handling background clutter, occlusion and appearance change.

scholarly journals Nonlinear joint model updating using static responses

2016 ◽  
Vol 8 (12) ◽  
pp. 168781401668265 ◽  
Author(s):  
Ping-Ping Yuan ◽  
Zuo-Cai Wang ◽  
Wei-Xin Ren ◽  
Xia Yang
Keyword(s):  
Finite Element ◽  
Finite Element Model ◽  
Objective Function ◽  
Model Updating ◽  
Nonlinear Behavior ◽  
Element Model ◽  
Joint Model ◽  
Joint System ◽  
Structural Joint ◽  
Nonlinear Joint

Nonlinear behavior is often observed in structural joint system due to external loads. A new technique of nonlinear structural joint model updating with static load test results is proposed in this article to investigate the actual behavior of a joint system. To calibrate the nonlinear parameters of the structural joint system, an appropriate finite element model is first established to characterize the complex nonlinear behavior caused by the joint connections. Combined with the sensitivity analysis, the parameters that describe the nonlinear behavior of the joint connections are selected as the parameters to be updated. Subsequently, an objective function is created in accordance with the residual between experimentally measured static deflections and analytically calculated static deflections through finite element model. The objective function is then optimized to obtain the proper values of the nonlinear force–displacement parameters with the regular simulated annealing algorithm. To validate the efficiency of this updating approach, two numerical examples under static concentrated loads are conducted. The obtained results indicate that the nonlinear joint model parameters can be successfully updated, and the updated new model can further forecast the true deflections of the nonlinear structure with good accuracy and stability.

scholarly journals A Narrow Deep Learning Assisted Visual Tracking with Joint Features

2020 ◽  
Vol 2020 ◽  
pp. 1-9
Author(s):  
Xiaoyan Qian ◽  
Daihao Zhang
Keyword(s):  
Visual Tracking ◽  
Model Updating ◽  
State Of The Art ◽  
Gaussian Mixture ◽  
Background Information ◽  
Robust Tracking ◽  
Data Set ◽  
Tracking Method ◽  
Benchmark Data ◽  
Updating Procedure

A robust tracking method is proposed for complex visual sequences. Different from time-consuming offline training in current deep tracking, we design a simple two-layer online learning network which fuses local convolution features and global handcrafted features together to give the robust representation for visual tracking. The target state estimation is modeled by an adaptive Gaussian mixture. The motion information is used to direct the distribution of the candidate samples effectively. And meanwhile, an adaptive scale selection is addressed to avoid bringing extra background information. A corresponding object template model updating procedure is developed to account for possible occlusion and minor change. Our tracking method has a light structure and performs favorably against several state-of-the-art methods in tracking challenging scenarios on the recent tracking benchmark data set.

Sign in / Sign up

Export Citation Format

Share Document