BangleFIR: bridging the gap in fashion image retrieval with a novel dataset of bangles

In this paper, we introduce Bangle Fashion Image Retrieval (BangleFIR), a novel dataset focusing on bangles within the fashion domain. While garment and footwear retrieval has garnered significant attention in recent years, intricate fashion items like jewellery have been comparatively overlooked due to their design complexity and the absence of suitable datasets. To bridge this gap, we have curated the BangleFIR dataset, comprising over 4.4k high-quality bangle images sourced from various online retailers and categorized into 56 distinct labels. Benchmarking the dataset with state-of-the-art image retrieval techniques demonstrated inadequate performance, concluding the need for dedicated research in this domain. We anticipate that the release of our dataset and benchmark will stimulate further exploration and innovation in the field. The dataset will be available to the public at: https://github.com/iammaidul/BangleFIR.