INSAIT study on AI systems recognised at international AI and computer vision forum

A study by the Institute of Computer Science, Artificial Intelligence and Technology (INSAIT) at Sofia University “St. Kliment Ohridski” has been recognised as a leading paper at one of the world’s most prestigious events in artificial intelligence and computer vision – the International Conference on Computer Vision, BTA reported.

The study's development, ObjectRelator, represents a significant step towards AI systems that perceive the world in a manner closer to human understanding. The model offers a novel approach to recognising and interpreting objects from any viewpoint, whether from the perspective of the human eye (the egocentric view) or from external cameras (the exocentric view).

While this ability comes naturally to humans, it remains a major challenge for current AI systems. Most leading computer vision models, such as SAM and SAM2, operate solely with images from a single viewpoint. ObjectRelator overcomes this limitation, opening the door to new applications in robotics, virtual and augmented reality, and the development of intelligent agents — AI systems capable of perceiving their environment, making decisions and acting autonomously to achieve tasks — that can transfer knowledge across different perspectives.

ObjectRelator is the result of a collaboration between researchers at INSAIT, the Swiss Federal Institute of Technology in Zurich, Fudan University in Shanghai, and the global technology company Ant Group, which serves over 1.3 billion users and 80 million merchants worldwide.