VPA - Computer Vision and Pattern Analysis Laboratory
Welcome to the Computer Vision and Pattern Analysis Laboratory of Sabanci University.
The Computer Vision and Pattern Analysis Lab of Sabanci University was established in 2001 by Prof. Aytül Erçil, and was selected as a "potential center of excellence" by the European Union. Its founding objective has been designing and implementing industrially applicable solutions using computer vision and pattern analysis techniques that are also theoretically fulfilling.
The members of VPALab work closely with local and foreign universities, governmental and private organizations to foster high-technology growth in the region. This growth is facilitated through substantive research, from theory to practical application and the training of skilled graduate and undergraduate engineers and scientists.
The main sources of research funding for VPALab include EU programs, EUREKA, NSF, government agencies and the industry. It is clearly recognized that many of the exciting new developments occurring in the convergent digital era require more integrated, collaborative and cross-disciplinary approaches. In recognition of this, the collaborative efforts at VPALab range not only cross discipline among various faculty members within the group, but also include other groups (e.g. Bioinformatics, Material Science, Mechatronics, Microelectronics programmes of Sabanci University) and other Universities (national and international).
For more information please check our Research and Publications pages.
Latest news from VPA lab
Mar. 2024 - 2 conference papers accepted to IGARSS'24 by Furkan Gül and Efkan Durakli, see you at Athens this July.
Feb. 2024 - New project on domain generalization for remote sensing with 1 PhD and 2 MSc fully funded positions (scholarship and accommodation) starting this spring. Contact Erchan Aptoula for details.
Jan. 2024 - Arın Zeyneloğlu has defended his MSc thesis on evidential deep learning
Dec. 2023 - Furkan Gül has defended his MSc thesis on a distance tranform based loss function for semantic segmentation
Dec. 2023 - Deren Ege Turan has defended her MSc thesis on explainable AI for remote sensing.
Nov. 2023 - The Spring 2024 Call of Graduate Applications will remain open until December 1st. Application Link
Oct. 2023 - There is a fully funded position for a MSc student on the topic of document segmentation and recognition for historical Ottoman documents. Contact B. Yanikoglu for details.
Oct. 2023 - There is a fully funded position for a MSc student on the topic of keyword spotting from historical handwritten Ottoman documents. Contact E. Aptoula for details.
Aug. 2023 - Figen Beken Fikri has defended her PhD thesis on deep learning based abstractive summarization.
Jul. 2023 - Mehmet Can Yavuz has defended his Ph.D. thesis on self-supervised approaches with biomedical and biometric application.
Jun. 2023 - E. Aptoula made a presentation about domain adaptation and domain generalization in the Machine Learning Summer School co-organized by Sabanci and Istanbul Technical Universities. Many thanks to the organizers Uğur Behçet Töreyin and Onur Varol for their invitation. Here are the slides (in Turkish) - link
Jun. 2023 - CFP for the Asian Conference on Machine Learning'23.
Jun. 2023 - E. Aptoula was kindly hosted by the School of Computer Science of the University of Lincoln, UK (Laboratory of Vision Engineering) where he provided a couple of presentations on remote sensing and deep domain adaptation. Many thanks to Petra Bosilj, Simon Parsons and Xujiong Ye - Presentation1 Presentation2.
May 2023 - A new event has been added SIU2023 - NST
Apr. 2023 - A webinar about the use of technology and remote sensing in particular for disaster management by B. Yanikoglu, I. Tekin and E. Aptoula (in Turkish) LINK
Jan. 2023 - E. Aptoula has been appointed as an Associated Editor for the IEEE Transactions on Geoscience and Remote Sensing journal.
Nov. 2022 - A new Tübitak 1001 project has been awarded to Prof. Berrin Yanikoglu on the "Automatic transcription of Ottoman Turkish documents through deep learning".
Nov. 2022 - Our joint paper with the University of Lincoln on domain generalization for object detection has just been accepted to AAAI'23.