• Generative Adversarial Networks

    It’s incredibly surprising to see super realistic, HD images of human faces, animals, and other things generated by an algorithm

  • Continual Learning

    Continual learning is the ability of a model to learn continually from a stream of data, building on what was learned previously

  • Segmentation

    Segmentation is the process of classifying each pixel belonging to a particular label



DeshuffleGAN enhances the learning of the discriminator and the generator, via a self-supervision approach

read more

Depth Estimation

Depth estimation is a pathway to understanding the 3D structure of a scene, which can be useful in a variety of tasks

read more

3D Point Cloud

A point cloud is a massive collection of points that express the characteristics of the surface of a target object

read more

Art, Architecture & AI

Can computers recognize visual design principles in an artistic composition?

read more

About ITU Vision Lab.

The development of computer vision research is faster than ever before. It has never attracted so society’s attention. We believe AI-based technology will be the main assistant of humanity in the future. Developing and deploying cutting-edge computer vision science to shape the changing world is one of the most important strategic decision for societies. Contributing to the development in progress is the main purpose of Vision Lab. Therefore, We —grad students, undergrads, postdocs— deploy cutting-edge scientific research and produce papers on Computer Vision.

read more

Exploring DeshuffleGANs in Self-Supervised Generative Adversarial Networks

G. Baykal, F. Özçelik and G. Ünal

A New Distributional Ranking Loss With Uncertainty: Illustrated in Relative Depth Estimation

A. Mertan, Y. H. Sahin, D. J. Duff and G. Unal

Diffusion MRI Spatial Super-Resolution Using Generative Adversarial Networks

E. Albay, U. Demir and G. Unal

Prof. Dr. Gözde Ünal


Faculty of Computer and Informatics Engineering, Room 5309
Ayazaga Campus, Istanbul Technical University
Istanbul, Turkey


Gozde Unal received her PhD in Electrical and Computer Engineering with a PhD minor in Mathematics from North Carolina State University, Raleigh, NC, USA, in 2002. After a postdoctoral fellowship at Georgia Institute of Technology, USA, Dr. Unal worked as a research scientist at Siemens Corporate Research, Princeton, NJ, USA between 2003-2007. She also worked at HPLabs, Palo Alto, CA, and Xerox Labs, Webster, NY as a visiting researcher.  She held positions of assistant professor and associate professor at Sabancı University, Faculty of Engineering and Natural Sciences between 2007-2015. In Fall 2015, she joined Istanbul Technical University, Department of Computer Engineering, where she is currently a full professor. She is one of the founding professors of the new AI&Data Engineering Department at ITU. She served as the founding director of the ITU-AI: ITU Artificial Intelligence and Data Science Research Center.

Dr. Unal was the recipient of L’Oreal Turkey’s Female Scientist Award in 2010 in Life Sciences, and the Distinguished Young Scientist Award from TUBA (GEBIP). She was awarded the Marie Curie Alumni Association (MCAA) Career Award of European Commission in 2017. She served as a technical program co-chair for MICCAI 2016, Athens, Greece, and MIDL 2019, London, UK. Currently, she is a member of the ITU-AI, and the director of the ITU Vision Lab. Her research interests are in AI, computer vision, deep learning and machine learning.

Doğay Kamar

Ph.D. Student

Fırat Öncel

Ph.D. Student

Gülçin Baykal Can

Ph.D. Student