BMCABodily Movement Collection and Annotation

Loading…

A high-quality and open-source data infrastructure designed with

the goal of computational understanding of human bodily expressions

Original

3D

A high-quality and open-source data infrastructure designed with the goal of computational understanding of human bodily expressions

Research on computational recognition of bodily expression in affective computing, AI, and HRI is progressing slowly because researchers often repeat the same work-intensive steps and lack large, high-quality, multifaceted datasets. We aim to provide the platform for high-quality, annotated, open-source, and extensible bodily-expression datasets.

Introducing the Bodily Movement Collection and Annotation (BMCA)

Handling an object

Sweeping

Clapping and opening the arms

Walking through a doorway

Standing from a sofa

Explaining with both hands

Turning and gesturing to the chest

Leaning in, then turning

Reaching across a work surface

Lifting a phone and tilting the head

Lowering an arm onto a fence

Walking and gesturing

Rising and stepping backward

Stepping forward with a raised arm

Reaching from a crouch

Lowering clasped hands and looking aside

A seated conversational gesture

Explaining with open palms

Reaching for a telephone

Saluting and bending forward

Raising an arm and pointing

Walking forward and reaching aside

Reaching while on the phone

Turning and lifting a cloth

Gesturing during a phone call

01–05 / 25

A growing library of

samples

Across 2 datasets

For the Community

Search through and filter samples from BMCA's dataset repository

Give to the community and contribute your dataset to BMCA

Download high-quality, annotated, and open-source datasets from BMCA's repository

Contributors

James Z. Wang

Penn State

Reg Adams

Penn State

Jia Li

Penn State

Sree Bhattacharyya

Penn State

Sitao Zhang

Advisory Board

Agata Lapedriza

Universitat Oberta de Catalunya

Alan F. Smeaton

Dublin City University

Stephen Wong

Houston Methodist

Gentiane Venture

University of Tokyo

Supported by

grant