DFC-YOLO: A Real-Time Product Label Detection Algorithm Based on YOLOv11 with Dual-Domain Selection and Chromaticity Pyramid

With the development of smart retail and smart shelf systems, real-time object detection of product labels has become one of the key technologies. Aiming at the problem that traditional detection methods have poor robustness in complex backgrounds and weak sensitivity to color information, this paper proposes an improved algorithm based on YOLOv11. The algorithm introduces a dual domain selection mechanism (DSM). By combining local detail perception with global spatial context modeling, it improves the model's ability to distinguish and express features in complex visual tasks, integrates image spatial domain and frequency domain features, and improves feature discrimination ability; at the same time, it designs a color gamut pyramid structure to extract color features of product labels at multiple scales and enhance the recognition ability of different color labels. Tested on a product label image dataset, the experiment shows that DFC-YOLO improves mAP50 by 1.5% over the baseline model. Compared with other models, DFCYOLO shows superior performance.