On the power of data augmentation for head pose estimation
This work addresses the lack of diverse training data for in-the-wild head pose estimation, offering incremental improvements for applications like human-computer interaction.
The paper tackles the problem of head pose estimation from monocular images by improving data augmentation and synthesis strategies, resulting in small, efficient models with very competitive accuracy for full 6 DoF pose estimation.
Deep learning has been impressively successful in the last decade in predicting human head poses from monocular images. However, for in-the-wild inputs the research community relies predominantly on a single training set, 300W-LP, of semisynthetic nature without many alternatives. This paper focuses on gradual extension and improvement of the data to explore the performance achievable with augmentation and synthesis strategies further. Modeling-wise a novel multitask head/loss design which includes uncertainty estimation is proposed. Overall, the thus obtained models are small, efficient, suitable for full 6 DoF pose estimation, and exhibit very competitive accuracy.