Geometric deep learning enables 3D kinematic profiling across species and environments.
Geometric deep learning enables 3D kinematic profiling across species and environments.
复制标题
DOI:
10.1038/s41592-021-01106-6
复制
发表时间:
2021-05
期刊:
影响因子:
48
通讯作者:
Ölveczky BP
中科院分区:
文献类型:
--
作者:
Dunn TW;Marshall JD;Severson KS;Aldarondo DE;Hildebrand DGC;Chettih SN;Wang WL;Gellis AJ;Carlson DE;Aronov D;Freiwald WA;Wang F;Ölveczky BP
Comprehensive descriptions of animal behavior require precise measurements of 3D whole-body movements. Although 2D approaches can track visible landmarks in restrictive environments, performance drops in freely moving animals, due to occlusions and appearance changes. Therefore, we designed DANNCE to robustly track anatomical landmarks in 3D across species and behaviors. DANNCE uses projective geometry to construct inputs to a convolutional neural network that leverages learned 3D geometric reasoning. We trained and benchmarked DANNCE using a 7-million frame dataset that relates color videos and rodent 3D poses. In rats and mice, DANNCE robustly tracked dozens of landmarks on the head, trunk, and limbs of freely moving animals in naturalistic settings. We extend DANNCE to datasets from rat pups, marmosets, and chickadees, and demonstrate quantitative profiling of behavioral lineage during development.
登录
查看更多内容
影响因子:
2.5
作者:
BOLLES, RC;WOODS, PJ
通讯作者:
WOODS, PJ
DOI:
10.1038/nrn.2015.8
发表时间:
2016-01
期刊:
Nature reviews. Neuroscience
影响因子:
--
作者:
Kalueff AV;Stewart AM;Song C;Berridge KC;Graybiel AM;Fentress JC
通讯作者:
Fentress JC
影响因子:
16.6
作者:
Bala, Praneet C.;Eisenreich, Benjamin R.;Zimmermann, Jan
通讯作者:
Zimmermann, Jan
影响因子:
16.6
作者:
Ayaz, Asli;Stauble, Andreas;Helmchen, Fritjof
通讯作者:
Helmchen, Fritjof
影响因子:
7.7
作者:
Machado, Ana S.;Darmohray, Dana M.;Carey, Megan R.
通讯作者:
Carey, Megan R.