Fish-Vista: A Multi-Purpose Dataset for Understanding & Identification of Traits from Images

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Mehrab, Kazi Sajeed, Maruf, M., Daw, Arka, Neog, Abhilash, Manogaran, Harish Babu, Khurana, Mridul, Feng, Zhenyang, Altintas, Bahadir, Bakis, Yasin, Campolongo, Elizabeth G, Thompson, Matthew J, Wang, Xiaojun, Lapp, Hilmar, Berger-Wolf, Tanya, Mabee, Paula, Bart, Henry, Chao, Wei-Lun, Dahdul, Wasila M, Karpatne, Anuj
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866916632360648704
author Mehrab, Kazi Sajeed
Maruf, M.
Daw, Arka
Neog, Abhilash
Manogaran, Harish Babu
Khurana, Mridul
Feng, Zhenyang
Altintas, Bahadir
Bakis, Yasin
Campolongo, Elizabeth G
Thompson, Matthew J
Wang, Xiaojun
Lapp, Hilmar
Berger-Wolf, Tanya
Mabee, Paula
Bart, Henry
Chao, Wei-Lun
Dahdul, Wasila M
Karpatne, Anuj
author_facet Mehrab, Kazi Sajeed
Maruf, M.
Daw, Arka
Neog, Abhilash
Manogaran, Harish Babu
Khurana, Mridul
Feng, Zhenyang
Altintas, Bahadir
Bakis, Yasin
Campolongo, Elizabeth G
Thompson, Matthew J
Wang, Xiaojun
Lapp, Hilmar
Berger-Wolf, Tanya
Mabee, Paula
Bart, Henry
Chao, Wei-Lun
Dahdul, Wasila M
Karpatne, Anuj
contents We introduce Fish-Visual Trait Analysis (Fish-Vista), the first organismal image dataset designed for the analysis of visual traits of aquatic species directly from images using problem formulations in computer vision. Fish-Vista contains 69,126 annotated images spanning 4,154 fish species, curated and organized to serve three downstream tasks of species classification, trait identification, and trait segmentation. Our work makes two key contributions. First, we perform a fully reproducible data processing pipeline to process images sourced from various museum collections. We annotate these images with carefully curated labels from biological databases and manual annotations to create an AI-ready dataset of visual traits, contributing to the advancement of AI in biodiversity science. Second, our proposed downstream tasks offer fertile grounds for novel computer vision research in addressing a variety of challenges such as long-tailed distributions, out-of-distribution generalization, learning with weak labels, explainable AI, and segmenting small objects. We benchmark the performance of several existing methods for our proposed tasks to expose future research opportunities in AI for biodiversity science problems involving visual traits.
format Preprint
id arxiv_https___arxiv_org_abs_2407_08027
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Fish-Vista: A Multi-Purpose Dataset for Understanding & Identification of Traits from Images
Mehrab, Kazi Sajeed
Maruf, M.
Daw, Arka
Neog, Abhilash
Manogaran, Harish Babu
Khurana, Mridul
Feng, Zhenyang
Altintas, Bahadir
Bakis, Yasin
Campolongo, Elizabeth G
Thompson, Matthew J
Wang, Xiaojun
Lapp, Hilmar
Berger-Wolf, Tanya
Mabee, Paula
Bart, Henry
Chao, Wei-Lun
Dahdul, Wasila M
Karpatne, Anuj
Computer Vision and Pattern Recognition
We introduce Fish-Visual Trait Analysis (Fish-Vista), the first organismal image dataset designed for the analysis of visual traits of aquatic species directly from images using problem formulations in computer vision. Fish-Vista contains 69,126 annotated images spanning 4,154 fish species, curated and organized to serve three downstream tasks of species classification, trait identification, and trait segmentation. Our work makes two key contributions. First, we perform a fully reproducible data processing pipeline to process images sourced from various museum collections. We annotate these images with carefully curated labels from biological databases and manual annotations to create an AI-ready dataset of visual traits, contributing to the advancement of AI in biodiversity science. Second, our proposed downstream tasks offer fertile grounds for novel computer vision research in addressing a variety of challenges such as long-tailed distributions, out-of-distribution generalization, learning with weak labels, explainable AI, and segmenting small objects. We benchmark the performance of several existing methods for our proposed tasks to expose future research opportunities in AI for biodiversity science problems involving visual traits.
title Fish-Vista: A Multi-Purpose Dataset for Understanding & Identification of Traits from Images
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2407.08027