Another Wavelength: Yuanyuan Sun

Sept. 8, 2021

This week in Another Wavelength, we chat with 3rd year Ph.D. Student Yuanyuan Sun. Yuanyuan is advised by Professor Rongguang Liang. 

Where are you from?

I come from a small town in the south part of China. 

What brought you to study optics?

I started to get interested in optics when I took the Optics course as an underground student majoring in physics. Then I joined an optics lab and did some computational algorithms as well as some optics experiments about polarimetry and spectrometers. I learned a lot and also had great fun. This experience encourages me to continue in optics.

Who is your hero in science?

Andrew Ng. Since I don't have a strong background in deep learning, I learned a lot from his online deep learning courses. But most importantly, I really like the concept behind the courses that education resources should be free and easily accessible for everyone no matter whether they are a student or not.

Describe your research in 20 words or fewer.

Hyperspectral recovery based on Deep learning.


Research caption: Network architecture of hyperspectral image recovery. Reference: Optics Letters 45.20 (2020): 5676-5679.

Describe your research in 200 words or fewer.

Hyperspectral imaging can capture both spatial (morphological) and spectral (chemical) information of the object which gives hyperspectral imaging broader application in biomedical imaging, remote sensing, face recognition, food inspection, earth observation, to name a few. However, commercial hyperspectral imaging devices are expensive and tend to suffer from the degradation of spatial , spectral or temporal resolution. To overcome these problems, hyperspectral image recovery using computational methods including deep learning is an affordable solution. We propose an end-to-end convolutional neural network to address such severely ill-posed problems and achieve great results.

Name three neat facts about you.

  1. I'm a great great great fan of DIY models kits. I can sit all day long building a model. (see two model's Yuanyuan created to the right)
  2. I am a ukulele learner right now.
  3. I enjoy cooking and baking.