"Brand Detection Dataset"

"This dataset supports object detection research focused on recognizing and localizing consumer product brands in retail environments. It contains images of five widely used brands, Colgate, Dettol, Classmate, Kellogg\u2019s, and Himalaya, depicting products both individually and in multi-brand arrangements. The images reflect realistic retail conditions such as shelf clutter, viewpoint variation, and lighting differences, enabling the development and evaluation of robust multi-class detection models.The dataset includes a combination of images captured manually by the authors in supermarket environments and a limited number of images obtained from publicly accessible online sources. All images are annotated with bounding boxes in COCO JSON format, ensuring compatibility with widely used computer vision frameworks such as PyTorch, TensorFlow, Detectron2, and MMDetection. Data augmentation techniques were applied to improve variability and generalization. The dataset is intended strictly for non-commercial research and educational purposes in areas such as brand detection, logo recognition, and retail analytics."