Person retrieval in surveillance using textual query: a review

Recent advancement of research in biometrics, computer vision, and natural language processing has discovered opportunities for person retrieval from surveillance videos using textual query. The prime objective a system is to locate description, e.g., short woman with pink t-shirt white skirt carrying black purse. She brown hair. Such description contains attributes like gender, height, type clothing, colour hair colour, accessories. are formally known as soft biometrics. They help bridge the semantic gap between human machine query person’s biometric attributes. It also not feasible manually search through huge volumes footage retrieve specific person. Hence, automatic vision language-based algorithms becoming popular. In comparison other state-of-the-art reviews, contribution paper follows: 1. Recommends most discriminative biometrics challenging conditions. 2. Integrates benchmark datasets methods performance evaluation. 3. A complete snapshot techniques based on features, classifiers, number attributes, deep neural networks, measures. 4. comprehensive coverage handcrafted features end-to-end approaches description.