Kang Zhao successfully passed his Master thesis exam on August 8th, 2019. His thesis was nominated to a best thesis award for Faculty of Graduate Studies by the thesis exam committee. Kang’s thesis is entitled with “Using deep neural networks for automatic building extraction with boundary regularization from satellite images”. Professors Costas Armenakis, Manos Pagelis and Mojgan Jadidi were served as Kang’s thesis exam committee. Kang has studied as new ways of integrating shape priors for predicting regular shapes of building outlines using satellite images. His research works will be published in conference and journal papers in collaboration with his colleagues in the lab (Andreas Wichmann, Mohammed Kamran and Jungwon Kang). His previous work was presented in CVPR Workshop (DeepGlobe Challenge) 2019. His work was ranked 4th in the DeepGlobe competition. Kang is looking for an opportunity to work in deep learning industry. We wish all the best with Kang’s bright future!