Research – CVML @ NUS

3D Pose and Shape Estimation

This research area is all about localizing the human body and hand joints in images and video. We target finding 2D and 3D keypoints as well as reconstructing the body and hand’s 3D surface, especially as a person interacts with other objects. Research outcomes will help downstream applications in human-computer interaction and augmented and virtual reality.

Human Pose Estimation [CVPR’23][TPAMI’23][ICLR’22][ICCV’21][IJCV’12][NeurIPS’11][ECCV’10]
Hand Pose Estimation [CVPR’23a][CVPR’23b][ICCV’21][BMVC’21][ECCV’20a][ECCV’20b][ICCV’19][CVPR’19a][CVPR’19b][CVPR’18][CVPR’17][ECCV’16]
3D Shape & Clothing [CVPR’22]

Video Understanding

This research area is all about the semantic interpretation of a video sequences. We specifically target videos with people to answer the fundamental question “What is the person doing?”. To that end, we tackle tasks such action recognition, temporal action segmentation, and action anticipation. Research outcomes will advance high-impact applications such as autonomous driving, home assistance robots, content-based video indexing and retrieval, and other intelligent visual systems.

Procedural Activity Understanding [CVPR’22][ICCV’19]
Temporal Action Segmentation [TMM’22][ECCV’22a][ECCV’22b][AAAI’22][CVPR’18]
Video Q&A [AAAI’22][CVPR’21]
Action Anticipation & Recognition[BMVC’21][ECCV’20][IJCV’12][PAMI’11][BMVC’11][CVPR’10]

Image & Video Enhancement

This research area is all about making images and videos look great!
We work on a wide range of tasks such as super-resolution, deblurring, and artifact removal.

Image Super-resolution [ECCV’22][GCPR’19]
Video Deblurring [CVPR’22]
Morphological Networks [PR’22][PR’20]

Others

Deep Regression [TPAMI’23][ICLR’23][ECCV’22][ICLR’22][ICCV’21]
Video & Image Segmentation [CVPR’22][BMVC’21][BMVC’20][GCPR’20][CVPR’19][GCPR’19]