Automated View and Path Planning for Scalable Multi-Object 3D Scanning (SIGGRAPH Asia 2016)
Xinyi Fan, Linguang Zhang, Benedict Brown, Szymon Rusinkiewicz
Robot In a Room: Toward Perfect Object Recognition in Closed Environments (arXiv:1507.02703)
Shuran Song, Linguang Zhang, Jianxiong Xiao
3d shapenets: A deep representation for volumetric shapes (CVPR 2015)
Zhirong Wu, Shuran Song, Aditya Khosla, Fisher Yu, Linguang Zhang, Xiaoou Tang, Jianxiong Xiao
Depth-Aware Shape from Shading (Undergraduate Final Year Thesis) Advisor: Professor Chi-Keung Tang
Traditional Shape from Shading (SfS) methods for single-view 3D reconstruction can usually achieve acceptable results, especially for high frequency components where normals frequently change. However, those methods often have erroneous results when handling low frequency components. Kinect depth map, though often noisy and in low quality, can ameliorate the errors in low frequency components when applying SfS. Our approach makes use of the depth map as guidance to upsample the height map reconstructed by SfS.
*Champion of the Best Final Year Thesis Award 2014
Ballbot Advisor: Professor Kam Tim Woo
My project partner fell in love with IMUs and he has a startup now! Check this out: sensabledynamics
In this project, we installed an IMU (MPU6050) to obtain the inclination of the robot. We used the first-order complementary filter to process the IMU data and PID controller to correct x-axis and y-axis inclination.