{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/robust-re-identification-of-manta-rays-from","title":"Robust Re-identification of Manta Rays from Natural Markings by Learning Pose Invariant Embeddings","arxiv_id":"1902.10847","date":"2019-02-28","proceeding":null,"authors":["Olga Moskvyak","Frederic Maire","Asia O. Armstrong","Feras Dayoub","Mahsa Baktashmotlagh"],"abstract":"Visual identification of individual animals that bear unique natural body\nmarkings is an important task in wildlife conservation. The photo databases of\nanimal markings grow larger and each new observation has to be matched against\nthousands of images. Existing photo-identification solutions have constraints\non image quality and appearance of the pattern of interest in the image. These\nconstraints limit the use of photos from citizen scientists. We present a novel\nsystem for visual re-identification based on unique natural markings that is\nrobust to occlusions, viewpoint and illumination changes. We adapt methods\ndeveloped for face re-identification and implement a deep convolutional neural\nnetwork (CNN) to learn embeddings for images of natural markings. The distance\nbetween the learned embedding points provides a dissimilarity measure between\nthe corresponding input images. The network is optimized using the triplet loss\nfunction and the online semi-hard triplet mining strategy. The proposed\nre-identification method is generic and not species specific. We evaluate the\nproposed system on image databases of manta ray belly patterns and humpback\nwhale flukes. To be of practical value and adopted by marine biologists, a\nre-identification system needs to have a top-10 accuracy of at least 95%. The\nproposed system achieves this performance standard.","url_abs":"http://arxiv.org/abs/1902.10847v1","url_pdf":"http://arxiv.org/pdf/1902.10847v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"robust-re-identification-of-manta-rays-from","repo_url":"https://github.com/olgamoskvyak/reid-manta","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":null,"task_name":"Triplet"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1902.10847","atlas_url":"https://app.syntology.ai/?focus=1902.10847","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}