BCSM-YOLO An Improved Product Package Recognition Algorithm for Automated

The increasing adoption of automation in retail stores has prompted the need for robust product recognition systems. This research proposes an enhanced product package recognition algorithm, BCSM-YOLO, designed to improve the accuracy and efficiency of object detection in automated retail environments. The algorithm leverages multiple versions of the YOLO (You Only Look Once) architecture, including YOLOv9, YOLOv10n, and YOLO26n, while introducing a novel variant, BCSM-YOLO10, to address challenges unique to retail scenarios. These challenges include varying lighting conditions, complex backgrounds, and a wide diversity of product packages, which can often lead to detection errors in traditional models.The proposed model is trained and evaluated using a comprehensive dataset sourced from Roboflow, which includes diverse retail products with different packaging types. BCSM-YOLO optimizes the existing YOLO architecture to enhance detection performance, achieving notable improvements in accuracy, speed, and real-time processing capabilities. The system is evaluated using key performance metrics such as mean average precision (mAP), recall, and inference time, demonstrating superior results compared to baseline models.This research contributes to the field of retail automation by offering a scalable and efficient product recognition solution. The BCSM-YOLO model not only improves inventory management and product detection but also enhances the overall customer experience by enabling fast, accurate recognition of products in dynamic retail environments. This work provides a significant step toward the widespread deployment of AI-powered systems in automated retail stores.

Keywords: Product Package Recognition, Automated Retail, YOLO, BCSM-YOLO, Object Detection, Computer Vision, Retail Automation, Deep Learning, Real-Time Detection, Roboflow Dataset.