An extended KCF tracking algorithm based on TLD structure in low frame rate videos

Moridvaisi, Hooman; Razzazi, Farbod; Pourmina, Mohammad Ali; Dousti, Massoud

doi:10.1007/s11042-020-08867-w

Cited by 8 publications

(3 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[3][4] 2010 CVPR, David S. Bolme [5] first applied correlation filtering to the field of tracking, and based on his idea, algorithms using correlation filtering for target tracking have appeared one after another, and the tracking effect has the tracking results are getting better and better. Moridvaisi et al [6] surmount KCF's limitations through the lens of the Tracking-Learning-Detection (TLD) framework and devised an algorithm that concurrently trains two classifiers, employing a semi-supervised co-training learning algorithm. Subsequently, they subject the proposed method to rigorous scrutiny against TB-100 datasets, juxtaposed with its counterparts.…”

Section: Introductionmentioning

confidence: 99%

Pedestrian Tracking Algorithm for Video Surveillance Based on Lightweight Convolutional Neural Network

Wei,

Zhai,

2024

IEEE Access

View full text Add to dashboard Cite

The Efficient Convolution Operators for Tracking (ECO) algorithm has garnered considerable attention in both academic research and practical applications due to its remarkable tracking efficacy, yielding exceptional accuracy and success rates in various challenging contexts. However, the ECO algorithm heavily relies on the deep learning Visual Geometry Group (VGG) network model, which entails complexity and substantial computational resources. Moreover, its performance tends to deteriorate in scenarios involving target occlusion, background clutter, and similar challenges. To tackle these issues, this study introduces a novel enhancement to the pedestrian tracking algorithm. Specifically, the VGG network is substituted with a lightweight MobileNet v2 model, thereby reducing computational demands. Additionally, a Double Attention Networks (A2-Net) module is incorporated to augment the extraction of crucial information, while pretraining techniques are integrated to expedite model convergence. Experimental results demonstrate that the C-ECO algorithm achieves comparable accuracy and success rates to the conventional ECO algorithm, despite reducing the model size by 27.96% and increasing the tracking frame rate by 46.11%. Notably, when compared to other prevalent tracking algorithms, the C-ECO algorithm exhibits an accuracy of 82.20% and a success rate of 64.72%. These findings underscore the enhanced adaptability of the C-ECO algorithm in complex environments, offering a more lightweight model while delivering superior tracking capabilities.

show abstract

Section: Introductionmentioning

confidence: 99%

Pedestrian Tracking Algorithm for Video Surveillance Based on Lightweight Convolutional Neural Network

Wei,

Zhai,

2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Among them, target positioning and tracking is the core of the intelligent milking robot, and the mechanical arm as the actuator of the automatic milking robot is the basis for realizing automatic milking. The algorithm is based on the KCF (Kernelized Correlation Filters) framework [16] , The KCF algorithm has achieved a good balance in terms of speed and accuracy [17] , but the original KCF algorithm has problems such as the inability to automatically identify targets and the scale cannot be adaptive [18] .To address the above problems, we improve the KCF algorithm, and the improved algorithm extracts the SURF (Speeded Up Robust Featur) [19] feature points of the target to determine the initial position of the target, and uses them as the target samples of the KCF algorithm to realize the localization of the cow teat region, and then obtains the target position in the world coordinate system by coordinate transformation. The coordinates are sent to the self-designed milking robotic arm platform to realize automatic milking operation.…”

Section: Introductionmentioning

confidence: 99%

Research on tracking algorithm of milking robot arm based on improved SURF-KCF

Zheng

Weng

et al. 2023

Second International Conference on Digital Society and Intelligent Systems (DSInS 2022)

View full text Add to dashboard Cite

Aiming at the slow movement and small displacement of dairy cows in the pasture breeding process, but the target scale changes significantly, a SURF-KCF dairy cow teat tracking algorithm is proposed. The original KCF algorithm needs to locate the target before tracking, and the tracking frame cannot be adaptive with the scale of the target. The SURF feature detection is introduced to provide target features for the KCF algorithm to achieve automatic target matching; and the scale estimation strategy is used to achieve scale adaption of cow teat targets. The experimental results show that the proposed improved algorithm achieves cow teat target matching and localization as well as scale adaption of the target area, and long-term stable tracking. Finally, in the laboratory environment, the improved algorithm is applied to the three-degree-of-freedom robotic arm automatic milking robot to realize the cow teat positioning and tracking experiment. When the cow's teat is detected, the robotic arm can automatically move to the position of the cow's teat. Good results have been obtained after testing, which shows the feasibility and effectiveness of the cow teat positioning system.

show abstract

“…After the model is established, the subsequent video sequences are searched to determine further whether the searched target or background is found or not. The common discriminative tracking algorithms include correlation filtering methods and deep learning methods [11][12][13]. Due to the success of the deep CNN in visual recognition tasks, a large number of studies have been performed using CNN for tracking algorithms [14,15].…”

Section: Introductionmentioning

confidence: 99%

A Vision‐Based Target Detection, Tracking, and Positioning Algorithm for Unmanned Aerial Vehicle

Liu

Zhang

2021

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

Unmanned aerial vehicles (UAV) play a pivotal role in the field of security owing to their flexibility, efficiency, and low cost. The realization of vehicle target detection, tracking, and positioning from the perspective of a UAV can effectively improve the efficiency of urban intelligent traffic monitoring. In this work, by fusing the target detection network, YOLO v4, with the detection-based multitarget tracking algorithm, DeepSORT, a method based on deep learning for automatic vehicle detection and tracking in urban environments, has been designed. With the aim of addressing the problem of UAV positioning a vehicle target, the state equation and measurement equation of the system have been constructed, and a particle filter based on interactive multimodel has been employed for realizing the state estimation of the maneuvering target in the nonlinear system. Results of the simulation show that the algorithm proposed in this work can detect and track vehicles automatically in urban environments. In addition, the particle filter algorithm based on an interactive multimodel significantly improves the performance of the UAV in terms of positioning the maneuvering targets, and this has good engineering application value.

show abstract

An extended KCF tracking algorithm based on TLD structure in low frame rate videos

Cited by 8 publications

References 32 publications

Pedestrian Tracking Algorithm for Video Surveillance Based on Lightweight Convolutional Neural Network

Pedestrian Tracking Algorithm for Video Surveillance Based on Lightweight Convolutional Neural Network

Research on tracking algorithm of milking robot arm based on improved SURF-KCF

A Vision‐Based Target Detection, Tracking, and Positioning Algorithm for Unmanned Aerial Vehicle

Contact Info

Product

Resources

About