Edvard Heikel

Papers

Total Citations

H-Index

About

Edvard Heikel is a researcher whose work lies at the intersection of computer vision, robotics, and semantic scene understanding. His primary research focus is on indoor scene recognition, a critical capability for social robots operating in human environments. Heikel’s major contribution is demonstrating that scene recognition can be performed effectively using only object-level information, challenging the field’s reliance on holistic image features. By applying a TF-IDF (Term Frequency-Inverse Document Frequency) weighting scheme—a technique borrowed from text retrieval—to detected objects, he showed that a robot can identify a room’s function (e.g., kitchen vs. office) purely from the objects it contains, without needing global visual context. This work, published in 2022, has already garnered 36 citations, signaling its impact on the growing field of semantic robotics. Heikel’s approach is particularly notable for its elegance and practicality: it leverages off-the-shelf object detectors and a simple, interpretable model, making it accessible for real-world robotic applications. His findings suggest that for social robots, understanding *what* is in a room may be more important than *how* the room looks, paving the way for more intuitive human-robot interaction.