Persistent animal identification leveraging non-visual markers: Tracking and Identification Dataset - Identifications Subset (TIDe-I)

This DataShare dataset pertains to Identification of group-housed mice as documented in the Thesis "Automated Identification and Behaviour Classification for Modelling Social Dynamics in Group-Housed Mice" (2023) by Michael Camilleri. It grew out of a collaboration with the Mary Lyon Centre at MRC Harwell, with the need to automatically identify group-housed mice using only position cues obtained from RFID tags. This sets the problem apart from the usual re-identification challenge, due to the mice have no visual markers to identify them. The challenge is compounded by the multiple mice which must be tracked/identified and the level of occlusion. We provide herein an annotated dataset containing mouse videos, pre-generated bounding-boxes and annotations of identifications at 4s intervals. The dataset can be used to train and evaluate identification methods. Further details are available at https://github.com/michael-camilleri/TIDe. The paper describing our work appears as: M. P. J. Camilleri, L. Zhang, R. S. Bains, A. Zisserman, and C. K. I. Williams, Persistent Animal Identification Leveraging Non-Visual Markers, CoRR (arXiv), cs.CV (2112.06809), Dec. 2021. [Available on arXiv](https://arxiv.org/pdf/2112.06809.pdf).