EnTimeMent
EU Horizon 2020 FET PROACTIVE project EnTimeMent (ENtrainment & synchronization at multiple TIME scales in the MENTal foundations of expressive gesture)
Pinned Repositories
Analysis-Qualities-of-Movement
The purpose of this software is to analyse Motion Capture data in order to calculate the joint with the maximum Shapley Value, separately for three motion features (speed, acceleration and angular momentum). The software models the human body as a weighted undirected graph, and calculates the importance of each joint in movement. A Lowpass Butterworth filter is used on the data for cleaning and processing. Each movement feature is calculated separately for each frame of each MoCap segment. Spectral clustering on the weighted graph is applied, to compute for each frame the Shapley Values of all the joints for all the movement features. It is a mixture of Python (for the data processing, filtering, motion feature calculation and results analysis) and MATLAB (for the creation of the body as a graph and for the Shapley Value calculation).
Ball-Exchange
TCN Neural Network architecture in the automatic detection of people drawing ellipses at multiple timescales
BANet
The code for a deep learning model BANet
Ellipses
TCN Neural Network architecture in the automatic detection of people drawing ellipses at multiple timescales
Graph-Generative-Flow
Hierarchical_HAR-PBD
huSync-DyadicSynchronization
onsetsync
Origin-of-Movement
The software, written in Matlab, extracts movement features (speed, tangential acceleration, kinetic energy, angular momentum) for two skeleton models of the human body, which refer to two different spatial scales. Then, such features are filtered and combined to compute a dissimilarity measure for the joints, from which a transferable utility game on an auxiliary graph is constructed. Finally, the vector of Shapley values for that game is computed (for the case in which the Shapley value coincides with the weighted degree centrality), and normalized with respect to the maximum Shapley value. The software allows for the possibility of computing the Kendall correlation between different rankings of joints. Further comments can be found inside the code.
Playing-Attention
Code embodying the P(l)aying Attention application.
EnTimeMent's Repositories
EnTimeMent/Analysis-Qualities-of-Movement
The purpose of this software is to analyse Motion Capture data in order to calculate the joint with the maximum Shapley Value, separately for three motion features (speed, acceleration and angular momentum). The software models the human body as a weighted undirected graph, and calculates the importance of each joint in movement. A Lowpass Butterworth filter is used on the data for cleaning and processing. Each movement feature is calculated separately for each frame of each MoCap segment. Spectral clustering on the weighted graph is applied, to compute for each frame the Shapley Values of all the joints for all the movement features. It is a mixture of Python (for the data processing, filtering, motion feature calculation and results analysis) and MATLAB (for the creation of the body as a graph and for the Shapley Value calculation).
EnTimeMent/Hierarchical_HAR-PBD
EnTimeMent/Origin-of-Movement
The software, written in Matlab, extracts movement features (speed, tangential acceleration, kinetic energy, angular momentum) for two skeleton models of the human body, which refer to two different spatial scales. Then, such features are filtered and combined to compute a dissimilarity measure for the joints, from which a transferable utility game on an auxiliary graph is constructed. Finally, the vector of Shapley values for that game is computed (for the case in which the Shapley value coincides with the weighted degree centrality), and normalized with respect to the maximum Shapley value. The software allows for the possibility of computing the Kendall correlation between different rankings of joints. Further comments can be found inside the code.
EnTimeMent/pstemr
This repository contains code for our P-STEMR (Parallel Space-Time Encoding Movement Representation) architecture that leverages action/activity recognition data for other movement interpretation tasks (e.g. pain level classification)
EnTimeMent/BANet
The code for a deep learning model BANet
EnTimeMent/Ellipses
TCN Neural Network architecture in the automatic detection of people drawing ellipses at multiple timescales
EnTimeMent/Graph-Generative-Flow
EnTimeMent/onsetsync
EnTimeMent/Ball-Exchange
TCN Neural Network architecture in the automatic detection of people drawing ellipses at multiple timescales
EnTimeMent/EnTimeMent.github.io
EnTimeMent/huSync-DyadicSynchronization
EnTimeMent/Playing-Attention
Code embodying the P(l)aying Attention application.
EnTimeMent/EmoPain_at_Home
Code for extracting data from the shared EmoPain@Home dataset
EnTimeMent/EMOSYNC_Behavioural_sync
Analysis of behavioural synchornisation metrics based on optical marker data during improvisation mirror game for a triad
EnTimeMent/EMOSYNC_EyeGaze
Eye gaze analysis based on the head marker data during improvisation mirror game for a triad
EnTimeMent/Group-Behavior-Recognition
EnTimeMent/Impact-of-Trajectory-Generation-Methods
EnTimeMent/MECS-Algorithm
EnTimeMent/MiMT
Neural network architecture for modelling affective movement behaviour at multiple timescales
EnTimeMent/raga-pose-estimation
EnTimeMent/Spook-and-Play
Analysis of synchronisation metrics for disembodied 4 people mirror game using Chronos interface