About Me
I am currently a second-year Master's student in Mechanical Engineering (Mechatronics track) at Stanford University. My work is centered around computer vision (low-level and deep learning), mechatronics and control and I am extremely interested in autonomous driving. Prior to joining Stanford, I graduated from Shanghai Jiao Tong University with highest honor. My resume can be downloaded [here].
Education
-
September 2012 - June 2016
Shanghai Jiao Tong University
Mechanical Engineering - Bachelor of Science (BS) summa cum laude
Shanghai - China
- Cumulative GPA: 90.39/100 (Rank: 1/184)
- Undergraduate thesis with Prof. Hong Shen in Experimental and Numerical Investigation of Laser Polishing.
- Chiang Chen Overseas Graduate Fellowship Awardee (10 nationwide per year)
Work Experience
-
June 2017 - September 2017
Autonomous Driving R&D Intern
Momenta.ai - Beijing - China
- Developed GTA V-based autonomous driving simulation platform in C++ for data collection and algorithm test.
- Wrote scripts in Python to clean raw data and generate ground-truth labels for training deep learning models.
- Trained and tested deep neural networks to predict car pose classification, wheel landing points and pillar locations.
- Calibrated game camera based on 2D and 3D correspondences extracted from GTA V.
-
September 2017 - Present
Course Assistant
Stanford University - Stanford, CA - USA
- Worked with Prof. Gordon Wetzstein on revising course materials of EE267: Virtual Reality.
Projects
-
Greedy Layerwise CNN Training for Weakly Supervised Object Localization
Z. Wang,C. Zhang
- Used global average pooling layer to preserve model’s localization ability and applied greedily layerwise training to improve its ability of extracting image features.
- Achieved 7% higher precision in classification and 20% better localization than benchmark.
-
OpenCV Lane Line Detection for Self-driving Cars
- Developed software pipeline to detect lane lines in images and videos for self-driving cars using OpenCV.
- Applied Gaussian smoothing to suppress noise and spurious gradients, ran Canny edge detector and Hough transform to obtain steady and continuous lane lines
-
End-to-end Scene Restoration and Object Removal
H. Li,C. Zhang
, C. Zhu
- Designed and proposed end-to-end pipeline for removing unwanted tourists in scenery pictures and videos
- Implemented person detector using pre-trained HOG in OpenCV with linear SVM to perform pedestrian detection on images and wrote scripts to generate mask for inpainting
-
Smart Fit
C. Zhang
, J. He, J.W. Suh, S. Juthacharoenwong
- Designed, built and tested Android-based fitting room assistance system with features of clothes identification, item request, personalized recommendations.
- Developed Android app with Google Firebase as backend interfacing with Arduino via Bluetooth and RFID.
-
Siege Tank Robot
J. He,C. Zhang
, Z. Shang, T. Lee
- Designed, built and tested an autonomous racing robot that navigates through a landscape with sensor system.
- Implemented hierarchal state machines and PID control with robust embedded code using event-driven framework, state charts, interrupts and SPI communication protocol
-
The Island: A First-person Tower Defense Game on Mobile VR
C. Zhang
, Q. Yu
- Developed first-person tower defense game for mobile VR for Google Cardboard using Unity and Google VR SDK.
- Built 3D terrain and designed main features, such as game logic, special effects, enemy AI, audio and animation.
- Implemented game features using Unity API in C# with teammate and deployed game on iOS devices.