Projects

 

My project experiences focus more on Computer Vision, Deep Learning and Robotics.

3D Face Reconstruction from RGB-D Images

12.2021 – 02.2022 ▪ Course Project of 3D Scanning & Motion Capture – with Daniel and Chang

My responsibilities in the team:

  • Designed the pipeline of the face reconstruction process and built up the code base in C++
  • Created the initial point cloud in 3D space
  • Researched on and studied facial landmark extraction and 3D Morphable Face Models
  • Performed coarse alignment with Basel Face Model using Procrustes Analysis
  • Optimized the fine optimization step to improve the performance of facial expression matching

3D Spatial Motion Estimation for Autonomous Driving

11.2020 – 10.2021 ▪ Guided Research Project (Master)

  • Estimated object motion by producing self-labelling data for self-supervised learning
  • Conducted experiments on computing 3D object motion labels with RGB and depth images
  • Designed a deep neural network to estimate 3D spatial motion

Novel View Synthesis for Self-Driving Cars

04.2020 – 10.2020 ▪ Practical Project (Master) – with Melis

* Note: the middle frame is the original input image while the other frames are all syntheses.

  • Predicted synthesized images at a new perspective view to make a stereo pair
  • Applied geometry theories and neural network architecture, U-Net, with self-supervised learning
  • Trained and tested on the KITTI benchmark and data generated by CARLA simulator
  • Outperformed most monocular sequences from KITTI on SLAM

Human Detection with Distance Estimation

09.2020 – 03.2021 ▪ Interdisciplinary Project (Master)

  • Applied existing framework to detect humans and predict their distance to the camera
  • Combined the modules on the robot prototype with a ROS node

Image Analyzer for People with Low Vision

09.2015 – 05.2016 ▪ Final Year Project (UG) ▪ Project Report – with Kaho & Joel

  • Built a real-time traveling aid with RGB-D camera to help the visually impaired travel safer and enlarge social circle via interaction with an audio-haptic feedback system
  • Mainly responsible for face & sign detection, recognition algorithms with encouraging results
  • Nominated for the Best FYP Award

Face Recognition in Smart Conference Room

06.2015 – 08.2015 ▪ Summer Research Project (UG Exchange Program in Germany)

  • Built a real-time face recognition system with OpenCV
  • Integrated to the room control system with text-to-speech to greet people getting in the room

Smart Car

01.2013 – 09.2014 ▪ “Freescale Cup” Intelligent Car Competition

  • Implemented a model car which could run on a track as fast as possible without control
  • Employed Linear CCD and other sensors to receive track images and analyze the condition to make automated response
  • Developed Embedded Software like pin test, motor control, PID control etc.
  • Developed the ColdFire / Kinetis MCU library and drivers with teammates
  • Sub-team leader of Optical Team 2 (Linear CCD) and attained Third Class award in 2014
  • Assisted in Smart Car Team which attained 3 awards in 2013

Electronic Piano with Virtual Keyboard

2014 ▪ Course Project of Embedded Systems (UG)

[cml_media_alt id='195']Electronic Piano with Virtual Keyboard [/cml_media_alt]

  • Built a system with a camera to identify fingers
  • Detect the pressed key via background cancellation, followed by loading WAV sound tracks from SD Card to STM32 via DMA & playing via DAC
  • Special features – able to change the piano mode to play various songs when pressing on keys