Visual is one of the main ways of human to perceive the world. Predicting Human Eye Fixation in visual content has been gaining lots of attention recently in images that have a different range of qualities and computer-rendered art effects images. I have been kindly given an opportunity to visit the University of Konstanz as a visiting researcher with Prof. Dr. Dietmar Saupe, Dr. Hanhe Lin on SFB-TRR 161 Project A05: Image/Video Quality Assessment.  It has been honouring to work with the team of the A05 visual lab.

I am a PhD student of the visual computing group at Cardiff University, UK. I was lucky  to get this visiting opportunity at the start of my PhD. My research topic focuses on Where Do People Look in Distorted Images, How is Attention Affected by Different Types and Different Levels of Distortion?, and build an automatic computational model to identify distortion-relevant salient regions so that Image Quality Assessment (IQA) can be performed based on this detection, as IQA algorithms fail in predicting perceived quality whereas humans can effortlessly judge the quality.

I had a wonderful time at the University of Konstanz, and I am fascinated by the landscape of the Bodensee. During the three-month research visit, I was engraved on the style of researching as it is different from the UK. All members of A05 vision lab closely work together with each other and collaborate in related areas with the lead of Prof. Dr. Dietmar Saupe. In the UK, we work more independently. I was gaining more experiences and pieces of knowledge through the group discussion of these experts in various visual tasks. It also left a deep impression on critical thinking ways on me.  Those three months truly gave me a new experience and helped me to build my knowledge from the expertise a lot. In the three months, I was not just getting new experiences on the subjective study, video quality assessment and crowdsourcing, but we also proposed A Statistical Evaluation of Visual Salience in Distortion-inducted Image Prediction Models to the ICIP conference meeting 2020. With Dr. Hanhe Lin, we proposed a new deep learning network in predicting human eye fixations when distortions occur in an image. So, two more papers are planned to be finished and get published in 2020. Those collaborative work will not just benefit from the findings of my previous research at Cardiff University supervised by Dr. Hantao Liu, and from the strong expertise in image quality and deep learning modelling at Prof. Dietmar Saupe’s lab.

Finally, it is hard to find words to express my gratitude. I want to say thank you to all the members of the group who gave me so many wonderful advises and became my friends: Dr. Vlad Hosu, Franz Götz-Hahn, Hui Men, Oliver Wiedemann. Also, I want to give special thanks to Claudia Widmann from SFB-TRR 161 who helped me to sort out my life in Konstanz,  Dr. Hanhe Lin who gave me big help in starting  my research,  Prof. Dr. Dietmar Saupe who led me in a right way to be a good researcher and my supervisor Dr. Hantao Liu who helped me to find this opportunity. Please accept my deepest thanks and I truly had a wonderful time in Konstanz.

Three Months in the Spirit of Collaborative Research

Xin Zhao is a PhD student at the School of Computer Science and Informatics of Cardiff University, UK. Her project is focused on eye attentive prediction and how Image Quality Assessment (IQA) could evaluate, monitor and optimise modern imaging systems.

Leave a Reply

Your email address will not be published.