Cargando…
A 3D virtual mouse generates synthetic training data for behavioral analysis
We develop a 3D synthetic animated mouse based on CT scans that is actuated using animation and semi-random, joint-constrained movements to generate synthetic behavioural data with ground-truth label locations. Image-domain translation produced realistic synthetic videos used to train 2D and 3D pose...
Autores principales: | , , , , , , , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
2021
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8034498/ https://www.ncbi.nlm.nih.gov/pubmed/33820989 http://dx.doi.org/10.1038/s41592-021-01103-9 |
_version_ | 1783676554609426432 |
---|---|
author | Bolaños, Luis A. Xiao, Dongsheng Ford, Nancy L. LeDue, Jeff M. Gupta, Pankaj K. Doebeli, Carlos Hu, Hao Rhodin, Helge Murphy, Timothy H. |
author_facet | Bolaños, Luis A. Xiao, Dongsheng Ford, Nancy L. LeDue, Jeff M. Gupta, Pankaj K. Doebeli, Carlos Hu, Hao Rhodin, Helge Murphy, Timothy H. |
author_sort | Bolaños, Luis A. |
collection | PubMed |
description | We develop a 3D synthetic animated mouse based on CT scans that is actuated using animation and semi-random, joint-constrained movements to generate synthetic behavioural data with ground-truth label locations. Image-domain translation produced realistic synthetic videos used to train 2D and 3D pose estimation models with accuracy similar to typical manual training datasets. The outputs from the 3D model-based pose estimation yielded better definition of behavioral clusters than 2D videos and may facilitate automated ethological classification. |
format | Online Article Text |
id | pubmed-8034498 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2021 |
record_format | MEDLINE/PubMed |
spelling | pubmed-80344982021-10-05 A 3D virtual mouse generates synthetic training data for behavioral analysis Bolaños, Luis A. Xiao, Dongsheng Ford, Nancy L. LeDue, Jeff M. Gupta, Pankaj K. Doebeli, Carlos Hu, Hao Rhodin, Helge Murphy, Timothy H. Nat Methods Article We develop a 3D synthetic animated mouse based on CT scans that is actuated using animation and semi-random, joint-constrained movements to generate synthetic behavioural data with ground-truth label locations. Image-domain translation produced realistic synthetic videos used to train 2D and 3D pose estimation models with accuracy similar to typical manual training datasets. The outputs from the 3D model-based pose estimation yielded better definition of behavioral clusters than 2D videos and may facilitate automated ethological classification. 2021-04-05 2021-04 /pmc/articles/PMC8034498/ /pubmed/33820989 http://dx.doi.org/10.1038/s41592-021-01103-9 Text en http://www.nature.com/authors/editorial_policies/license.html#termsUsers may view, print, copy, and download text and data-mine the content in such documents, for the purposes of academic research, subject always to the full Conditions of use: http://www.nature.com/authors/editorial_policies/license.html#terms |
spellingShingle | Article Bolaños, Luis A. Xiao, Dongsheng Ford, Nancy L. LeDue, Jeff M. Gupta, Pankaj K. Doebeli, Carlos Hu, Hao Rhodin, Helge Murphy, Timothy H. A 3D virtual mouse generates synthetic training data for behavioral analysis |
title | A 3D virtual mouse generates synthetic training data for behavioral analysis |
title_full | A 3D virtual mouse generates synthetic training data for behavioral analysis |
title_fullStr | A 3D virtual mouse generates synthetic training data for behavioral analysis |
title_full_unstemmed | A 3D virtual mouse generates synthetic training data for behavioral analysis |
title_short | A 3D virtual mouse generates synthetic training data for behavioral analysis |
title_sort | 3d virtual mouse generates synthetic training data for behavioral analysis |
topic | Article |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8034498/ https://www.ncbi.nlm.nih.gov/pubmed/33820989 http://dx.doi.org/10.1038/s41592-021-01103-9 |
work_keys_str_mv | AT bolanosluisa a3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT xiaodongsheng a3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT fordnancyl a3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT leduejeffm a3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT guptapankajk a3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT doebelicarlos a3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT huhao a3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT rhodinhelge a3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT murphytimothyh a3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT bolanosluisa 3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT xiaodongsheng 3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT fordnancyl 3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT leduejeffm 3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT guptapankajk 3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT doebelicarlos 3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT huhao 3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT rhodinhelge 3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis AT murphytimothyh 3dvirtualmousegeneratessynthetictrainingdataforbehavioralanalysis |