FENGTING YANG

Ph.D. Candidate · Computer Vision and Deep Learning · fuy34 [at] psu.edu

I am a fourth-year Ph.D. student supervised by Dr. Zihan Zhou. Before joining Penn State, I worked with Dr. Bin Wu , and received my Master degree in Instrument Science and Technology and Bachelor degree in Measuring and Control Technology and Instrumentations from Tianjin University in 2017 and 2014, respectively.

My research interest lies on computer vision and deep learning, especially focusing on 3D vision. Currently, I am working on projects that use deep learning methods to recover 3D structures from image(s). Please refer to my CV for more details.


PUBLICATIONS

Superpixel Segmentation with Fully Convolutional Networks

Fengting Yang, Qian Sun, Hailin Jin, and Zihan Zhou
In IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2020 [pdf] [supp] [code]

Recovering 3D Planes from a Single Image via Convolutional Neural Networks

Fengting Yang and Zihan Zhou
In European Conference on Computer Vision (ECCV). 2018 [pdf] [supp] [code]

A Cost-effective Non-orthogonal 3D Measurement System

Fengting Yang, Bin Wu, Ting Xue, Mohammed Farhan Ahmed and Jie Huang
Measurement, 128 (2018), pp.264-270. [pdf]

The Error Analysis of the Non-Orthogonal Total Station Coordinate Measurement System

Bin Wu, Wen Ding, Fengting Yang, and Ting Xue
Acta Metrologica Sinica, 2017, 38(6): 661-666 (In Chinese) [pdf]

A Novel Calibration Method for Non-orthogonal Shaft Laser Theodolite Measurement System

Bin Wu, Fengting Yang, Wen Ding, and Ting Xue.
Review of Scientific Instruments 87, no. 3 (2016): 035102. [pdf]

PATENTS

A Non-orthogonal Laser Total Station Based 3D Coordinate Measurement Method

Bin Wu, Ting Xue, and Fengting Yang
China Invention Patent, ZL 201610915794, May 2019 [link]

An Inverse Kinematic Model for Non-orthogonal Laser Theodolite

Ting Xue, Bin Wu, and Fengting Yang
China Invention Patent, ZL 201610949270, Mar. 2019 [link]

EXPERIENCE

PROFESSIONAL EXPERIENCE

Amazon

Applied Scientist Intern

- Wrote software for a novel camera device, collected data with it, and performed data pre-processing.
- Explored multiple depth estimation methods for various inputs and cues, and delivered the model ahead of the deadline;
- The final model outperformed the baseline by 32% and exceed the targeted accuracy by 21% in terms of RMSE;
- Delivered the report, and advised the team on how to apply the proposed technologies on Amazon's physical shopping products.

May. 2020 - Aug. 2020

ByteDance (TikTok)

Reserch Intern

- Developed a weakly supervised single-view plane recovery method that can simultaneously predict plane segments and parameters with only depth supervision.
- The final model is able to handle arbitrary number of planes, and improved parameter accuracy by 27% compared to PlaneRecover with the same plane number.

May. 2019 - Aug. 2019

Zhejiang Nuoke Electronic Technology Development Co., Ltd

Research Intern

- Designed a ZigBee based ad-hoc mesh network for real-time outdoor communication.

Feb. 2014 - May. 2014

TEACHING EXPERIENCE

DS340 - Applied Data Science

Teaching Assistant

College of Information Scienece and Techonolgy, The Pennsylvania State University

Aug. 2020 - Dec. 2020

Principle of Automatic Control

Teaching Assistant

School of Precision Instruments and Optoelectronics Engineering, Tianjin University

Sept. 2015 - Feb. 2016

AWARDS

  • National Scholarship for Graduate Students, China, 2016 (2%)
  • Chiang Chen Scholarship, Chiang Chen Industrial Charity Foundation, 2014, 2015 (5%)
  • First-Class Academic Scholarship, Tianjin University, 2014 (10%)
  • Outstanding Graduate, Tianjin University, 2014
  • Third Prize in iCAN Contest (Tianjin Division), China International NEMS Societ, 2012

Miscellaneous

Apart from being a Ph.D. student, I enjoy my time with movies, computer games, and travel. I also like to play badminton and billiards, regardless of my poor skill.

I am from a small city called Wenzhou in China. I am very pround of my hometown and our dialect, Wenzhou Hua. Some people rank it as the world's most difficult dialect and call it "Devil's lanuage" . Here is an interesting comparison between my hometown language and Cantonese . Have fun!