Laundry Detergent Occlusion Image Dataset

The retail e-commerce sector faces significant challenges due to the increasing competition and the need for effective product recognition under various conditions. Existing solutions often struggle with the accurate detection of products when they are partially occluded, leading to poor customer experiences. This dataset aims to address the technical issue of object detection under occlusion conditions, specifically for laundry detergent products like Tide, Persil, Ariel, and All. The dataset is structured to facilitate the training and evaluation of models in adversarial scenarios. Data is collected using high-resolution cameras in controlled lighting environments, ensuring consistency and quality. Quality control measures include multi-round annotations, consistency checks, and expert reviews to maintain high standards. Data is organized in a JPEG format with structured metadata for easy access and usage.