Title:	Horse 3D reconstruction
DNr:	Berzelius-2023-49
Project Type:	LiU Berzelius
Principal Investigator:	Ci Li <cil@kth.se>
Affiliation:	Kungliga Tekniska högskolan
Duration:	2023-03-08 – 2023-10-01
Classification:	10299
Keywords:

Abstract

Horses are essential animals for various domains, such as sports, agriculture, transportation, and therapy. Understanding their anatomy, physiology, and behavior is critical to ensuring their health, welfare, and performance. One promising approach to enhance this understanding is through 3D reconstruction, which can create detailed and accurate models of horses' body shapes and movements. However, 3D horse reconstruction requires significant manual effort and expertise, limiting its scalability and reproducibility. In this proposal, we aim to address this challenge by developing a deep learning and computer vision technique for 3D horse reconstruction. We will train and validate our technique using existing datasets of horse images and videos, as well as new data that we will collect from various sources, such as veterinary clinics, farms, and competitions. Besides, we will also investigate other modality for reconstruction such as audio.