2021/05/12 by Avi Ben-Cohen, Nadav Zamir, Ben-Cohen, Avi +9 · 1 citation
Computer Science · #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Text and Document Classification Technologies #cs.CV
paper · pdf · doi:10.48550/arxiv.2105.05926
arxiv created 2021/05/12 · openalex publication_date 2021/05/12 · arxiv updated 2021/05/14 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Training a neural network model for recognizing multiple labels associated with an image, including identifying unseen labels, is challenging, especially for images that portray numerous semantically diverse labels. As challenging as this task is, it is an essential task to tackle since it represents many real-world cases, such as image retrieval of natural images. We argue that using a single embedding vector to represent an image, as commonly practiced, is not sufficient to rank both relevant seen and unseen labels accurately. This study introduces an end-to-end model training for multi-label zero-shot learning that supports semantic diversity of the images and labels. We propose to use an embedding matrix having principal embedding vectors trained using a tailored loss function. In addition, during training, we suggest up-weighting in the loss function image samples presenting higher semantic diversity to encourage the diversity of the embedding matrix. Extensive experiments show that our proposed method improves the zero-shot model's quality in tag-based image retrieval achieving SoTA results on several common datasets (NUS-Wide, COCO, Open Images).