Vehicle Tracking based on an Improved DeepSORT Algorithm and the YOLOv4 Framework

Perera, Imalie; Senavirathna, Shehan; Jayarathne, Aseni; Egodawela, S.M.K.C.S.B.; Godaliyadda, G. M. R. I.; Ekanayake, Parakrama; Wijayakulasooriya, J. V.; Herath, Vijitha; Sathyaprasad, Sathindra

doi:10.1109/iciafs52090.2021.9606052

Cited by 10 publications

(6 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It can detect every vehicle, even at a small size, in any weather condition, including heavy snow, fog, rain, or dust. Perera et al [20] implemented vehicle tracking and detection to reduce errors in tracking associations. Kapania et al [21] proposed a multi-object detection and tracking method for Unmanned Aerial Vehicle (UAV) datasets and achieved a Multi-Object Tracking Accuracy (MOTA) of 45.8%.…”

Section: A Object Detection and Tracking Methodsmentioning

confidence: 99%

An Improved Speed Estimation Using Deep Homography Transformation Regression Network on Monocular Videos

et al. 2023

View full text Add to dashboard Cite

Vehicle speed estimation is one of the most critical issues in intelligent transportation system (ITS) research, while defining distance and identifying direction have become an inseparable part of vehicle speed estimation. Despite the success of traditional and deep learning approaches in estimating vehicle speed, the high cost of deploying hardware devices to get all related sensor data, such as infrared/ultrasonic devices, Global Positioning Systems (GPS), Light Detection and Ranging (LiDAR systems), and magnetic devices, has become the key barrier to improvement in previous studies. In this paper, our proposed model consists of two main components: 1) a vehicle detection and tracking componentthis module is designed for creating reliable detection and tracking every specific object without doing calibration; 2) homography transformation regression networkthis module has a function to solve occlusion issues and estimate vehicle speed accurately and efficiently. Experimental results on two datasets show that the proposed method outperforms the state-of-the-art methods by reducing the mean square error (MSE) metric from 14.02 to 6.56 based on deep learning approaches. We have announced our test code and model on GitHub with https://github.com/ervinyo/Speed-Estimation-Using-Homography-Transformation-and-Regression-Network.

show abstract

Section: A Object Detection and Tracking Methodsmentioning

confidence: 99%

An Improved Speed Estimation Using Deep Homography Transformation Regression Network on Monocular Videos

et al. 2023

View full text Add to dashboard Cite

show abstract

“…DeepSORT integrated a Convolution Neural Network (CNN) model for generating features used as a deep association metric. According to the research of Perera et al [39], DeepSORT's feature extractor is not suitable for vehicles since the CNN model was trained using a large-scale person reidentification dataset [40]. With this, instead of using the existing network of DeepSORT that was trained for person reidentification, AlexNet was used by Perera et al [39] to extract the appearance features of each detected object or vehicle.…”

Section: B Vehicle Trackingmentioning

confidence: 99%

“…According to the research of Perera et al [39], DeepSORT's feature extractor is not suitable for vehicles since the CNN model was trained using a large-scale person reidentification dataset [40]. With this, instead of using the existing network of DeepSORT that was trained for person reidentification, AlexNet was used by Perera et al [39] to extract the appearance features of each detected object or vehicle. This study extracted the appearance feature with a size of 4096×1 from the second FCN layer of AlexNet for the implementation of OCSORT-DA.…”

Section: B Vehicle Trackingmentioning

confidence: 99%

Observation-Centric with Appearance Metric for Computer Vision-Based Vehicle Counting

Brillantes,

Sybingco,

Billones

et al. 2023

JAIT

View full text Add to dashboard Cite

Tracking objects in video sequences is a key step for applications involving computer vision like traffic monitoring and security systems. Occlusion is a frequent problem in object tracking which can result in the tracker losing track of the occluded object or misidentifying it with the occluding object. Moreover, the limited memory and computing power of traffic analysis systems presents a scaling problem, especially in object tracking applications. This paper aims to improve object tracking performance by minimizing data association errors in low frame rate tracking applications. Reducing frame rates alleviates memory and computing power limitations, and utilizing a tracker that can handle occlusion can address occlusion-related issues in object tracking. The proposed tracking method, Mask-OCSORT, uses the observation-tracking method with cosine similarity, intersection-over-union, and velocity consistency metrics for the association problem. The paper analyzes the effect of using bounding box and mask predictions of deep learning models in generating tracks. This study uses evaluation metrics like HOTA, MOTA, and IDF1 to assess the proposed tracking method, and employs evaluation metrics such as precision, recall, and F-score to assess the counting based on generated IDs from the tracking method. The study applied the Mask-OCSORT tracking for vehicle counting application and achieved an F-score of 87.18% at 5 frames per second (fps), and 75% at 1 fps.

show abstract

“…SORT uses a combination of motion and appearance features to track objects and can handle occlusions, clutter, and other challenges associated with object tracking. SORT is a simple yet effective algorithm that can track multiple objects simultaneously and is widely used in various applications such as surveillance [17] and self-driving cars [18]. The algorithm is computationally efficient, making it suitable for real-time applications, and it has a high tracking accuracy and robustness, even in cluttered environments.…”

Section: Drone Trackingmentioning

confidence: 99%

Real-Time Drone Detection and Tracking in Distorted Infrared Images

Bentamou,

Zein-Eddine,

Messai

et al. 2023

2023 IEEE International Conference on Image Processing Challenges and Workshops (ICIPCW)

View full text Add to dashboard Cite

With the increasing use of drones for various applications, their detection and tracking have become critical for ensuring safety and security. In this paper, we propose an algorithm for detecting and tracking drones from infrared (IR) images in challenging conditions such as noise and distortion. Our algorithm involves YOLOv7 for drone detection and utilizes the SORT algorithm for real-time tracking. To detect distortion in the drone images, we employed a vision transformer in parallel with a customized CNN. The experimental results demonstrate the effectiveness of our approach in challenging conditions and highlight the potential for future developments in drone detection and tracking using deep learning techniques. We achieve a precision of 94.2%, a recall of 92.64%, and a mean average precision (mAP) of 92.6% on the provided test data. The implementation code can be found at: https://github.com/a-bentamou/Drone-detectionand-tracking.

show abstract

Vehicle Tracking based on an Improved DeepSORT Algorithm and the YOLOv4 Framework

Cited by 10 publications

References 9 publications

An Improved Speed Estimation Using Deep Homography Transformation Regression Network on Monocular Videos

An Improved Speed Estimation Using Deep Homography Transformation Regression Network on Monocular Videos

Observation-Centric with Appearance Metric for Computer Vision-Based Vehicle Counting

Real-Time Drone Detection and Tracking in Distorted Infrared Images

Contact Info

Product

Resources

About