Robust Component Analysis for Facial Analysis and Recognition in the Wild

该项目的目标是开发从高维时间数据中无监督提取低维表示的方法，能够处理“野外”数据。上述方法的应用可能包括：-基于运动数据的人验证/识别-人类行为的时空对齐和无监督分割-基于视觉数据的人类行为相似性的无监督测量

The objective of this project is the development of methods for unsupervised extraction of low-dimensional representations from high-dimensional temporal data, capable of handling "in the wild" data.Application of the aforementioned methods may include:-Person verification/recognition based on motion data-Spatial and temporal alignment and unsupervised segmentation of human behaviour-Unsupervised measurement of human behavior similarity based on visual data