Aleksis Pirinen

Papers

Total Citations

H-Index

About

Aleksis Pirinen is a researcher whose work lies at the intersection of computer vision, robotics, and embodied AI. His primary contributions focus on developing intelligent agents that can actively learn from their environments, particularly through the lens of semantic scene understanding. In his highly cited work, "Embodied Visual Active Learning for Semantic Segmentation" (2021, 26 citations), Pirinen tackles a fundamental challenge: how can an autonomous agent efficiently explore a 3D environment to build a robust visual understanding by strategically selecting which views to request human annotation? This research bridges the gap between passive deep learning benchmarks and the real-world need for sample-efficient, interactive perception. By pioneering methods for active view selection, Pirinen’s work has significant implications for deploying vision systems in robotics, autonomous navigation, and augmented reality, where data is scarce and exploration is costly. His research pushes beyond static datasets, advocating for agents that learn to see by deciding where to look—a critical step toward truly autonomous visual intelligence.