Dataset Information

A three-dimensional virtual mouse generates synthetic training data for behavioral analysis.

ABSTRACT: We developed a three-dimensional (3D) synthetic animated mouse based on computed tomography scans that is actuated using animation and semirandom, joint-constrained movements to generate synthetic behavioral data with ground-truth label locations. Image-domain translation produced realistic synthetic videos used to train two-dimensional (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.

SUBMITTER: Bolanos LA

PROVIDER: S-EPMC8034498 | biostudies-literature | 2021 Apr

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A three-dimensional virtual mouse generates synthetic training data for behavioral analysis.

Bolaños Luis A LA Xiao Dongsheng D Ford Nancy L NL LeDue Jeff M JM Gupta Pankaj K PK Doebeli Carlos C Hu Hao H Rhodin Helge H Murphy Timothy H TH

Nature methods 20210405 4

We developed a three-dimensional (3D) synthetic animated mouse based on computed tomography scans that is actuated using animation and semirandom, joint-constrained movements to generate synthetic behavioral data with ground-truth label locations. Image-domain translation produced realistic synthetic videos used to train two-dimensional (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 ...[more]

PMID: 33820989

Similar Datasets

Project description:Facial appearance transfer from donor to recipient in face transplantation is a concern. Previous studies of facial appearance transfer and facial appearance persistence (preservation of the recipient's facial likeness) in face transplants simulated using two-dimensional photographic manipulations found low facial appearance transfer (2.6 percent) and high facial appearance persistence (66 percent). Three-dimensional computer simulation of complex facial transplant patterns may improve the accuracy of facial appearance transfer and facial appearance persistence estimations.Three-dimensional virtual models of human faces were generated from deidentified computed tomographic angiographs and used as "donors" or "recipients" for virtual face transplantation. Surgical planning software was used to perform 73 virtual face transplantations by creating specific facial defects (mandibular, midface, or large) in the recipient models and restoring them with allografts extracted from the donor models. Twenty independent reviewers evaluated the resemblance of each resulting posttransplant model to the donor (facial appearance transfer) and recipient (facial appearance persistence). The results were analyzed using tests for equal results with one-sample and pairwise Rao-Scott Pearson chi-square testing, correcting for clustering and multiple testing.Overall rates of facial appearance persistence and facial appearance transfer were high (69.2 percent) and low (32.4 percent), respectively. The mandibular pattern had the highest rates of facial appearance persistence and lowest rates of facial appearance transfer. Facial appearance persistence and transfer were similar across sexes.Facial appearance persistence is high and facial appearance transfer is low after virtual face transplantation. Appearance transfer and persistence after virtual face transplantation are more dependent on the anatomy than on the size of transplanted facial aesthetic units. This information may reassure recipients of partial face transplants and donor families.

Dataset Information

A three-dimensional virtual mouse generates synthetic training data for behavioral analysis.

Publications

A three-dimensional virtual mouse generates synthetic training data for behavioral analysis.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets