CVMI Lab @ HKU

3D Scene Understanding

We aim at developing technologies for representation learning with 3D data (e.g. point clouds); RGB-D/3D semantic understanding (e.g. semantic segmentation, object detection, and instance segmentation) and registration tasks.

Data-efficient and Computation-efficient Deep Learning

We are interested in developing computation-efficent deep learning algorithms and data-efficient deep learning methods including high-quality data synthesis methods, efficient image annotation systems, label-efficient learning strategies (e.g. self-supervised learning, weakly supervised learning, domain adaptation, and semi-supervised learning, etc).

High-quality 3D Reconstruction

We are dedicated to pushing the boundaries of 3D sensing with learning-based approaches, e.g. 3D object & scene reconstruction from a single image or videos; and sparse-to-dense or depth completion to enhance depth-sensing quality.

3D Scene Understanding

We aim at developing technologies for representation learning with 3D data (e.g. point clouds); RGB-D/3D semantic understanding (e.g. semantic segmentation, object detection, and instance segmentation) and registration tasks.

Data-efficient and Computation-efficient Deep Learning

We are interested in developing computation-efficent deep learning algorithms and data-efficient deep learning methods including high-quality data synthesis methods, efficient image annotation systems, label-efficient learning strategies (e.g. self-supervised learning, weakly supervised learning, domain adaptation, and semi-supervised learning, etc).

High-quality 3D Reconstruction

We are dedicated to pushing the boundaries of 3D sensing with learning-based approaches, e.g. 3D object & scene reconstruction from a single image or videos; and sparse-to-dense or depth completion to enhance depth-sensing quality.

CVMI Lab

Computer Vision and Machine Intelligence
The University of Hong Kong

3D Scene Understanding

Data-efficient and Computation-efficient Deep Learning

High-quality 3D Reconstruction

3D Scene Understanding

Data-efficient and Computation-efficient Deep Learning

High-quality 3D Reconstruction

News

News

Our Research

Our Research

CVMI Lab

Computer Vision and Machine Intelligence The University of Hong Kong

3D Scene Understanding

Data-efficient and Computation-efficient Deep Learning

High-quality 3D Reconstruction

3D Scene Understanding

Data-efficient and Computation-efficient Deep Learning

High-quality 3D Reconstruction

News

News

Our Research

Our Research

Computer Vision and Machine Intelligence
The University of Hong Kong