Linearly augmented real-time 4D expressional face capture

Personalised 3D face creation has always been a hot topic in the computer vision community. Many methods have proposed including statistic model, non-rigid registration and high-end depth acquisition equipment. However, practical applications, those existing still their own limitations. For example, performance of model-based highly depends on generality pre-trained model; based are sensitive to quality input data; equipment-based less able be popularised due expensive equipment costs; deep learning-based can only perform well if proper training data provided for target domain, require GPU better performance. To this end, paper presents an adaptive template augmented method that automatically obtain personalised 4D facial modelling using consumer-grade device. The noisy from such cheap device handled. whole process consists series linear solutions achieved real-time online processing CPU computation laptop. There is no constraint nor complex operation required by method. No additional time-consumptive pre- or post-processing personalisation needed. Comparisons against several demonstrate superiority