Research
3D Information Retrieval from 2D Face Images
We developed an algorithm that is able to retrieve 3D geometric information of human faces from 2D images/videos. The applications of this algorithm include
Head pose estimation (estimation of roll, yaw, pitch angles)
3D face detection (detection of 3D bounding boxes)
Virtual glasses try on
See below videos for demonstration:
Our method can run in real-time on ordinary laptops without the need of GPU or depth cameras.
Automatic Parking Using Deep Reinforcement Learning (parking lots, parallel parking, presentation)
3D Face Reconstruction from a Single Image
C.-P. Wei and Y.-C. F. Wang, "With One Look: 3D Face Shape Estimation From a Single Snapshot," IEEE International Conference on Multimedia & Expo (ICME), July 2016 (oral presentation, acceptance rate: 15%). [pdf, slides, flowchart]
Undersampled Face Recognition via Robust Auxiliary Dictionary Learning
C.-P. Wei and Y.-C. F. Wang, "Undersampled Face Recognition via Robust Auxiliary Dictionary Learning," IEEE Transactions on Image Processing, vol. 24, no. 6, pp. 1722-1734, 2015. [pdf, code]
C.-P. Wei and Y.-C. F. Wang, "Undersampled Face Recognition with One-Pass Dictionary Learning," IEEE International Conference on Multimedia & Expo (ICME), July 2015 (acceptance rate: 30%). [pdf, poster]