Applying Machine Learning and Computer Vision for Planogram Compliance Evaluation in Retail Environments

This research investigates the limitations of manual approaches in evaluating planogram compliance in retail settings and proposes an automated solution leveraging machine learning and computer vision. The primary objective is to design a system capable of accurately and efficiently assessing planogram compliance. The methodology encompasses the collection and preprocessing of real-world image data from retail stores in Vietnam, followed by detailed product annotation using the Roboflow platform. Subsequently, object detection models—YOLOv11 and YOLOv12—are trained for product recognition and classification tasks. Detected products are grouped into rows via the DBSCAN clustering algorithm, and layout alignment with predefined planograms is performed using the Hungarian algorithm. Experimental evaluations indicate that the proposed system achieves high precision in product detection and offers reliable planogram compliance assessment. The findings highlight the system's potential to reduce human error, improve shelf space utilization, and enhance operational efficiency in retail environments.