Robust Histogram Signature-Driven Template Matching for Vehicle Tracking in Traffic Videos

Abstract

Correlation-based template matching (CTM) is widely used for object detection because of its simplicity and effectiveness in scenarios where grayscale features are sufficient. However, CTM often underperforms when color is a crucial distinguishing factor. To address this limitation, we propose contextual hierarchical composite template matching (CH-CTM), a color-histogram-enhanced CTM algorithm that integrates color information into a traditional correlation framework. CH-CTM augments the correlation index with red, green, and blue histogram comparisons to improve robustness in color-sensitive contexts. We evaluated CH-CTM using five diverse traffic video datasets that include various lighting conditions, vehicle types, sizes, and colors. Twelve experiments were conducted using standard performance metrics. Results demonstrated notable improvements over baseline CTM: CH-CTM achieved a peak accuracy of 98.30%, an average accuracy of 92.43%, and average precision of 92%. These findings confirm the importance of incorporating color information into template matching, which expands CTM’s applicability in complex real-world scenarios, particularly in traffic surveillance and object tracking.

Received: 17 March 2025 | Revised: 30 April 2025 | Accepted: 19 May 2025

Conflicts of Interest

The author declares that he has no conflicts of interest to this work.

Data Availability Statement

Data are available from the corresponding author upon reasonable request.

Author Contribution Statement

Aaron Rasheed Rababaah: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration.