AIG-Dataset

The Apparel Images Gallery (AIG) dataset is a well-curated collection of 130,000 color images representing a wide variety of clothing items, designed to support advanced research in fashion analysis and computer vision. Each image is uniformly resized to a resolution of 224×224×3, ensuring consistency for deep learning model training. The dataset encompasses 13 apparel categories—blouses, coats, dresses, hoodies, jackets_vests, jeans, pants, shirts, shorts, skirts, sweaters, trousers, and T-shirts—with 10,000 images per class, resulting in a balanced and diverse dataset. These categories are also grouped by garment type, with 7 classified as top-wear and 6 as lower-wear. Covering extensive apparel attributes such as style, length, texture, and color variations, the AIG dataset serves as a rich resource for tasks like fashion item classification, clothing attribute prediction, and content-based image retrieval in fashion-related applications.