Research Article
A Machine Learning Approach to Detect Violent Behaviour from Video
@INPROCEEDINGS{10.1007/978-3-030-16447-8_9, author={David Nova and Andr\^{e} Ferreira and Paulo Cortez}, title={A Machine Learning Approach to Detect Violent Behaviour from Video}, proceedings={Intelligent Technologies for Interactive Entertainment. 10th EAI International Conference, INTETAIN 2018, Guimar\"{a}es, Portugal, November 21-23, 2018, Proceedings}, proceedings_a={INTETAIN}, year={2019}, month={4}, keywords={Machine learning Support Vector Machine Action recognition Pose estimation Video analysis}, doi={10.1007/978-3-030-16447-8_9} }- David Nova
André Ferreira
Paulo Cortez
Year: 2019
A Machine Learning Approach to Detect Violent Behaviour from Video
INTETAIN
Springer
DOI: 10.1007/978-3-030-16447-8_9
Abstract
The automatic classification of violent actions performed by two or more persons is an important task for both societal and scientific purposes. In this paper, we propose a machine learning approach, based a Support Vector Machine (SVM), to detect if a human action, captured on a video, is or not violent. Using a pose estimation algorithm, we focus mostly on feature engineering, to generate the SVM inputs. In particular, we hand-engineered a set of input features based on keypoints (angles, velocity and contact detection) and used them, under distinct combinations, to study their effect on violent behavior recognition from video. Overall, an excellent classification was achieved by the best performing SVM model, which used keypoints, angles and contact features computed over a 60 frame image input range.