however, began during his undergraduate years at International Institute of Information Technology Hyderabad. He was selected by IAPR as a Next Generation Young Researcher to share his experience.
Discovering Research at IIIT-H
Rohan pursued his Bachelor’s degree in Computer Science and Engineering at IIIT Hyderabad, where the institute’s Honours Research Program gave students an opportunity to engage in independent research under faculty mentorship.
Having developed a strong interest in mathematics and machine learning during his coursework, Rohan decided to step into research during his third year. This decision led him to the Center for Visual Information Technology (CVIT), where he began working under the guidance of Ravi Kiran Sarvadevabhatla.
“At the beginning, I had very limited exposure to deep learning and research methodologies,” Rohan recalls. “But the experience of working on a real-world research problem pushed me to learn continuously and think critically.”
Building Foundations in Pattern Recognition
His research focused on textual attribute recognition in scanned multilingual documents—specifically detecting attributes such as bold, italic, underline, and strikeout for individual words in document images. The larger goal was to improve large-scale document digitization systems.
To tackle the problem, Rohan first strengthened his understanding of computer vision, neural networks, and deep learning fundamentals. At the same time, he explored classical image processing approaches to understand the problem from multiple perspectives.
This process became his introduction to the broader field of pattern recognition.
“Through this work, I started understanding how models learn from data, how data representation affects predictions, and how AI systems can be interpreted and improved,” he says.
As his research progressed, he also explored Optical Character Recognition (OCR) systems and document layout analysis models, gaining insight into how AI can process and understand complex document structures.
Recognition at ICDAR 2025
A major milestone in Rohan’s research journey came when his work received recognition at International Conference on Document Analysis and Recognition 2025, one of the leading international conferences in the field of document analysis and pattern recognition.
His paper, which addressed challenges in textual attribute recognition for scanned multilingual documents, received the Best Paper Award, highlighting both the novelty of the research and its practical relevance to large-scale document digitization and OCR systems.
The recognition marked an important achievement in his undergraduate research journey and reinforced the impact of combining strong fundamentals with applied problem-solving.
Perspectives on Supporting Young Researchers
Having worked across multiple institutions and research groups—from document analysis at IIIT-H to medical imaging at IIT Delhi and generative modeling at IISc—Rohan believes that some of the most meaningful learning experiences happen beyond the boundaries of a single lab.
According to him, organizations such as the International Association for Pattern Recognition can play a significant role in supporting young researchers by creating opportunities for broader engagement and collaboration.
“One of the most transformative learning experiences for young researchers comes from interacting with people working on diverse but related problems,” he says.
Rohan strongly advocates for expanding funded summer schools, describing them as more than just educational events. For many students, especially those from developing regions, such initiatives provide their first meaningful exposure to the global research community. By reducing financial barriers, these programs help ensure that talented researchers can access mentorship, advanced learning, and collaborative opportunities regardless of background.
He also highlights the need for more structured research communities within organizations like IAPR. Rather than interactions occurring only around conferences and paper deadlines, he envisions domain-specific groups that bring together senior researchers, postdoctoral scholars, and students working in related areas.
Such collaborative communities, he believes, could become continuous support systems where researchers discuss open problems, exchange ideas, and learn from one another in a more accessible and sustained manner.
“Some of the most valuable discussions I’ve had were with people working on adjacent problems. A structured research community can make those interactions more frequent and significantly reduce the struggles faced by young researchers,” he reflects.
Applying Research in Industry
Today, Rohan applies many of these foundational learnings in the industry, working on practical AI systems that solve real-world problems using computer vision technologies.
His journey reflects how undergraduate research experiences can shape careers, build confidence, and open pathways into advanced AI and pattern recognition research.
For students exploring research opportunities, Rohan’s story serves as a reminder that stepping into unfamiliar problem spaces, asking questions, and engaging with the broader research community can often become the starting point for meaningful innovation and growth.