VPALab, as proven by its track record, maintains a sustained flow of research support. The general areas of expertise in the lab have evolved around the following topics:
Biomedical Image Analysis (Skin lesion classification, cell recognition, microscopy image analysis, etc)
Biometrics (Face recognition, voice recognition, fingerprint recognition, signature recognition, hand shape recognition, multimodal biometrics)
Handwriting Recognition (Online, offline, English, Turkish, Ottoman)
Image and Video understanding (Faces, plants, crowds, vehicles, etc)
Intelligent Manufacturing (Invariant object recognition, web inspection, automated tolerance inspection, surface quality inspection, etc.)
Medical Image Analysis (Covid-19 detection, Alzheimer's detection, MRI analysis, x-ray analysis, etc)
Precision Agriculture (Fruit counting, weed/crop detection, etc)
Remote Sensing (Pixel & scene classification, content-based retrieval, hyperspectral and multi-modal analysis, water-quality estimation, etc.)
Domain generalized remote sensing image analysis (2024-26)
This project (Tübitak 1001) is a collaboration between Dr. E. Aptoula (SU) and Dr. Koray Kayabol from Gebze Technical University. It addresses the domain shift in common remote sensing image analysis tasks, such as scene classification, object detection and change detection. It aims to design and implement novel domain generalization solutions adapted specifically for remote sensing multi-modal data, coming from multispectral and synthetic aperture radar sensors. Contact Dr. E. Aptoula for details.
Visual keyword spotting (2023-25)
This project (funded by Sabanci University) is a collaboration between Dr. M. Kuru from the History department, and Dr. E. Aptoula and the objective is to develop deep learning based solutions that will enable the retrieval of visual keywords from scanned archives of Ottoman handrwritten documents. The Ottoman archive consists of hundreds of millions of samples, where classic OCR techniques are in general challenging to apply due to calligraphic writing styles. We thus aim to provide a valuable tool for the history research community.
We have developed a deep learning system to detect Covid-19 in CT scans (Tübitak 1001). Our deep learning ensemble is able to detect Covid-19 infection with 90% accuracy, from both healthy images as well as other lung infections. The system is now deployed and being used at İstanbul University Cerrahpaşa Schol of Medicine Hospital, with 95% reported accuracy. This is a joint work with Sabanci University, Istanbul University Cerrahpaşa School of Medicine and TÜBİTAK UME. Contact Berrin Yanikoğlu for more details.
In this project (Tübitak 1001), we work on different aspects of Face Attribute classification, ranging from building state-of-art deep learning systems to understand binary or relative attributes and to semi-supervised learning in the same domain. Graduate students: Sara Atito Ali Ahmed & Mehmet Can Yavuz. Contact Berrin Yanikoğlu for more details.
Hate speech detection
In recent years, disseminating hate speech around the world has increased significantly. To counter this ever-growing threat to society, many automated systems to detect hate speech in social media have been recently proposed by the research community. This project (Tübitak Bilateral Collaboration), aims to develop hate speech detection systems in Turkish and Arabic separately, taking into account the specificity and richness of the morphological structure inherent in each language. Contact Berrin Yanikoğlu for more details.
Intelligent Public Safety Platform for Smart Cities (2021-24)
A Türkiye-Qatar project (Tübitak Bilateral Collaboration) realized jointly with Havelsan, Turkcell, Hyperion, IstLink, Tazi, Ozyegin University., Texas A&M University, Hamad Bin Khalifa University and Informatica Qatar. We design and develop algorithms and solutions for crowd counting and crowd behaviour analysis from optical video feeds. Contact Erchan Aptoula for more details.
Bingöl honey pollen recognition (2019-22)
Bingöl is renounced for its honey, and has won the first place multiple times in international contests. In this project the focus is on its pollen-wise and genetic branding. Our role is to design and develop a computer vision system for pollen localization and identification from optical microscopy images. This project is a collaboration between Sabanci, Gebze Technical and Bingöl Üniversities. . Contact Erchan Aptoula for more details.
MULTI-variate, temporal, resolution and SourCe remote sensing image Analysis and Learning (MULTISCALE) (2019-22)
A Türkiye-France project (Tübitak Bilateral Collaboration) jointly with Laetitia Chapel from the Université de Bretagne Sud, France. MULTISCALE is a research project that aims at providing a complete and integrated framework for multiscale image analysis and learning with hierarchical representations of complex remote sensing images. Contact Erchan Aptoula for more details.