RM-Depth: Unsupervised Learning of Recurrent Monocular Depth in Dynamic Scenes

1 code implementation • CVPR 2022 • Tak-Wai Hui

In this paper, an unsupervised learning framework is proposed to jointly predict monocular depth and complete 3D motion including the motions of moving objects and camera.

Monocular Depth Estimation Unsupervised Monocular Depth Estimation

LiteFlowNet3: Resolving Correspondence Ambiguity for More Accurate Optical Flow Estimation

1 code implementation • ECCV 2020 • Tak-Wai Hui, Chen Change Loy

The keys to success lie in the use of cost volume and coarse-to-fine flow inference.

Optical Flow Estimation Scene Flow Estimation

Inter-Region Affinity Distillation for Road Marking Segmentation

1 code implementation • CVPR 2020 • Yuenan Hou, Zheng Ma, Chunxiao Liu, Tak-Wai Hui, Chen Change Loy

We study the problem of distilling knowledge from a large deep teacher network to a much smaller student network for the task of road marking segmentation.

Knowledge Distillation Lane Detection +1

Learning to Synthesize Fashion Textures

no code implementations • 18 Nov 2019 • Wu Shi, Tak-Wai Hui, Ziwei Liu, Dahua Lin, Chen Change Loy

Another important observation is that fashion textures are multi-modal.

A Lightweight Optical Flow CNN -- Revisiting Data Fidelity and Regularization

3 code implementations • 15 Mar 2019 • Tak-Wai Hui, Xiaoou Tang, Chen Change Loy

Over four decades, the majority addresses the problem of optical flow estimation using variational methods.

Optical Flow Estimation

LiteFlowNet: A Lightweight Convolutional Neural Network for Optical Flow Estimation

4 code implementations • CVPR 2018 • Tak-Wai Hui, Xiaoou Tang, Chen Change Loy

FlowNet2, the state-of-the-art convolutional neural network (CNN) for optical flow estimation, requires over 160M parameters to achieve accurate flow estimation.

Optical Flow Estimation

Motion-Depth: RGB-D Depth Map Enhancement with Motion and Depth in Complement

no code implementations • CVPR 2014 • Tak-Wai Hui, King Ngi Ngan

Since the spatial resolution of the color image is generally higher than that of the depth image, this paper introduces a new method to enhance the depth images captured by a moving RGB-D system using the depth cues from the induced optical flow.

3D Reconstruction Image Enhancement +1

Determining Motion Directly from Normal Flows Upon the Use of a Spherical Eye Platform

no code implementations • CVPR 2013 • Tak-Wai Hui, Ronald Chung

We address the problem of recovering camera motion from video data, which does not require the establishment of feature correspondences or computation of optical flows but from normal flows directly.

Motion Estimation Translation