Erdi Sarıtaş, a researcher at Istanbul Technical University, took part as a researcher in an ambitious research project focusing on one of the most challenging technological problems of our time in his scientific career. In the project titled "Facial Recognition from Low-Resolution Images", he developed innovative solutions to the problem of accurate identification with low-quality images, which is one of the weakest links in security systems. In this process, he had the opportunity to deepen his researcher identity by taking responsibility in both theoretical modeling and practical application areas.

Within the scope of the project, three different strategies were developed to solve the matching problem between low-resolution (LR) and high-resolution (HR) facial images. Sarıtaş worked on a hybrid model that combines reducing the quality of gallery images, increasing the quality of query images with the "facial hallucination" method, and these two approaches. In this experimental process, he was able to significantly improve the accuracy of facial recognition systems by applying multi-scale and multi-hypothesis methods.

"In real-world scenarios, verifying an identity, especially from a blurry image from security cameras, was like digging a well with a needle. Facing this challenging problem took my scientific curiosity and algorithmic thinking skills to the next level. The process of combining different strategies to create a hybrid solution solidified my identity as a researcher between theoretical knowledge and practical results. The education I received at ITU was very effective in developing an analytical and creative approach to this technological problem I faced," says Sarıtaş.

The project's scientific outputs have not only resonated with academic circles but have also garnered international acclaim. The results obtained were first presented at a prestigious conference such as the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2023, and then expanded and published in the IEEE Access 2025 journal. This success proved that the solutions offered by the project are both innovative and feasible.

While ITU produces solutions to critical problems in future technologies such as artificial intelligence and image processing, it also offers our young researchers like Erdi Sarıtaş the opportunity to take firm steps forward in the scientific world and create social benefit. This experience of Sarıtaş is the product of an environment where our students gain not only knowledge but also a culture of scientific production and innovation.

Erdi Sarıtaş, a researcher at Istanbul Technical University, took part as a researcher in an ambitious research project focusing on one of the most challenging technological problems of our time in his scientific career. In the project titled "Facial Recognition from Low-Resolution Images", he developed innovative solutions to the problem of accurate identification with low-quality images, which is one of the weakest links in security systems. In this process, he had the opportunity to deepen his researcher identity by taking responsibility in both theoretical modeling and practical application areas.

Within the scope of the project, three different strategies were developed to solve the matching problem between low-resolution (LR) and high-resolution (HR) facial images. Sarıtaş worked on a hybrid model that combines reducing the quality of gallery images, increasing the quality of query images with the "facial hallucination" method, and these two approaches. In this experimental process, he was able to significantly improve the accuracy of facial recognition systems by applying multi-scale and multi-hypothesis methods.

"In real-world scenarios, verifying an identity, especially from a blurry image from security cameras, was like digging a well with a needle. Facing this challenging problem took my scientific curiosity and algorithmic thinking skills to the next level. The process of combining different strategies to create a hybrid solution solidified my identity as a researcher between theoretical knowledge and practical results. The education I received at ITU was very effective in developing an analytical and creative approach to this technological problem I faced," says Sarıtaş.

The project's scientific outputs have not only resonated with academic circles but have also garnered international acclaim. The results obtained were first presented at a prestigious conference such as the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2023, and then expanded and published in the IEEE Access 2025 journal. This success proved that the solutions offered by the project are both innovative and feasible.

While ITU produces solutions to critical problems in future technologies such as artificial intelligence and image processing, it also offers our young researchers like Erdi Sarıtaş the opportunity to take firm steps forward in the scientific world and create social benefit. This experience of Sarıtaş is the product of an environment where our students gain not only knowledge but also a culture of scientific production and innovation.

Erdi Sarıtaş, a researcher at Istanbul Technical University, took part as a researcher in an ambitious research project focusing on one of the most challenging technological problems of our time in his scientific career. In the project titled "Facial Recognition from Low-Resolution Images", he developed innovative solutions to the problem of accurate identification with low-quality images, which is one of the weakest links in security systems. In this process, he had the opportunity to deepen his researcher identity by taking responsibility in both theoretical modeling and practical application areas.

Within the scope of the project, three different strategies were developed to solve the matching problem between low-resolution (LR) and high-resolution (HR) facial images. Sarıtaş worked on a hybrid model that combines reducing the quality of gallery images, increasing the quality of query images with the "facial hallucination" method, and these two approaches. In this experimental process, he was able to significantly improve the accuracy of facial recognition systems by applying multi-scale and multi-hypothesis methods.

"In real-world scenarios, verifying an identity, especially from a blurry image from security cameras, was like digging a well with a needle. Facing this challenging problem took my scientific curiosity and algorithmic thinking skills to the next level. The process of combining different strategies to create a hybrid solution solidified my identity as a researcher between theoretical knowledge and practical results. The education I received at ITU was very effective in developing an analytical and creative approach to this technological problem I faced," says Sarıtaş.

The project's scientific outputs have not only resonated with academic circles but have also garnered international acclaim. The results obtained were first presented at a prestigious conference such as the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2023, and then expanded and published in the IEEE Access 2025 journal. This success proved that the solutions offered by the project are both innovative and feasible.

While ITU produces solutions to critical problems in future technologies such as artificial intelligence and image processing, it also offers our young researchers like Erdi Sarıtaş the opportunity to take firm steps forward in the scientific world and create social benefit. This experience of Sarıtaş is the product of an environment where our students gain not only knowledge but also a culture of scientific production and innovation.