Cheng Zhang

Graduate Student

UESTC

Biography

I am a third-year Master’s student of machine learning and computer vision under the supervision of Prof. Shuaicheng LIU at University of Electronic Science and Technology of China. During my master study, I worked closely with Prof. Zhaopeng Cui and Dr. Yinda Zhang. My research interests lie in the intersection of Computer Vision, Machine Learning, and Computer Graphics. Currently, I am working on 3D scene understanding, 3D reconstruction, and 3D detection. I received my Bachelor’s degree at UESTC in 2019.

Interests

3D reconstruction
3D detection
3D scene understanding

Education

MEng in Information and Communication Engineering, 2022

University of Electronic Science and Technology of China
BEng in Internet of Things Engineering, 2019

University of Electronic Science and Technology of China

Publication

Cheng Zhang, Zhaopeng Cui, Cai Chen, Shuaicheng Liu, Bing Zeng, Hujun Bao, Yinda Zhang

August 2021 International Conference on Computer Vision (ICCV)

DeepPanoContext: Panoramic 3D Scene Understanding with Holistic Scene Context Graph and Relation-based Optimization

We propose a novel method for panoramic 3D scene understanding which recovers the 3D room layout and the shape, pose, position, and semantic category for each object from a single full-view panorama image.

Cheng Zhang, Zhaopeng Cui, Yinda Zhang, Bing Zeng, Marc Pollefeys, Shuaicheng Liu

March 2021 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

Holistic 3D Scene Understanding from a Single Image with Implicit Representation

We present a new pipeline which takes a single image as input, estimates layout and object poses, then reconstructs the scene with Signed Distance Function (SDF) representation.

Projects

Deep Learning Server System

A deep learning system for small research teams, with multiple GPU servers, centralized network storage, server and user management, 10 GbE network, and per-server SSD cache.

FMCW Positioning Radar

A FMCW Radar with 4 transmitting antennas and 12 receive antennas working within the 2.7 GHz - 3.7 GHz band that can locate a person in the same room with accuracy of 20 cm.

Ball on Big Plate

The project focus on balancing and controlling the movement of a metal ball on a 65cm wide plate, and is a work of 3 team-mates which get us first prize in 2017NUEDC.

Ball on Small Plate

The project focus on balancing and controlling the movement of a metal ball on a 20cm wide plate, and is capable of setting target coordinate, circle around the center with adjustable radius.

Ball on Beam

Balancing a ball on a beam with a Raspberry Zero, a camera, one servo, a STM32F103 board.

Rotary Inverted Pendulum

A rotary inverted pendulum system built with an encoder, a brush motor, a STM32F103 board that is capable of setting target orientation, non-stop swinging and automatically swinging up.

Recent Posts

Aug 17, 2019 1 min read Photoblog

Long Exposure of Airplane Turning

The conditions and timing for taking this photo were demanding. First, the city lights in the backgound is only visible in a full left turn. Second, measures must be taken to avoid the shaking of hand and the reflection on the window.

Nov 29, 2015 1 min read Photoblog

Ginkgo of UESTC

Both campuses of UESTC in Chengdu are famous for their ginkgo. When autumn is in full swing, there is a constant stream of visitors.

Jul 13, 2015 1 min read Photoblog

Milky Way at Qinhai Lake

In Milky Way Photography, Japanese photography Masahiro Miyasaka created a method to capture close-range objects (e.g. flowers) and stars far away in a single shot.