article
Detecting moving objects is a vital task within intelligent transportation systems and diverse computer vision applications. A resilient, semi-automatic method provides unsupervised object detection by pairing background subtraction with a tailored version of the YOLOv8 neural network. The process begins by applying low-rank decomposition for background subtraction to identify moving targets. A clustering algorithm subsequently refines these segmented outputs. These refined detections serve as training data to fine-tune the modified YOLOv8 model, effectively creating an automated pipeline that removes the requirement for manual dataset labelling. The system performs both object detection and classification. Benchmark evaluations conducted on the UA-DETRAC and CDnet 2014 datasets show that this approach achieves notable improvements in mean Average Precision when compared to recent alternative techniques.
Manual labelling of image datasets is costly, slow, and labour-intensive. By combining classical background subtraction with modern neural networks, this approach automates dataset creation and fine-tunes detection models without human intervention. This makes training computer vision systems for moving object detection more efficient and scalable across real-world environments such as transport infrastructure.
The technique is directly applicable to intelligent transportation systems and video monitoring operations, such as traffic surveillance and flow tracking. Prospective users include transport authorities, city planners, and developers of video analytics software. The technology appears to be applied and tested, having been validated against public benchmarks, but operational field trials would be required to advance it toward commercial deployment.
AI-generated from the published abstract. Always read the original work before citing.
Detecting moving objects is a pivotal research do-main within Intelligent Transportation Systems (ITS) and various computer vision applications. This paper presents an innovative and resilient semi-automatic method for unsupervised object detection, which combines a customized adaptation of the CNN-based You Only Look Once V8 (YOLOv8) approach with background subtraction. First, background subtraction using low-rank decomposition is employed to detect moving objects. Next, a clustering method is utilized to refine the outcomes of the background subtraction process. These refined results are then employed to fine-tune the modified YOLOv8, which is subsequently used for object detection and classification. The primary contribution of this work lies in its novel detection framework that eliminates the need for manual labeling by creating an automatic labeling system. Extensive experiments carried out on practical object monitoring benchmarks reveal that the proposed framework notably enhances mean Average Precision (mAP) when compared with recent approaches on both the UA-DETRAC and CDnet 2014 datasets.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.1109/niles63360.2024.10753164
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
