Inamoto Kouya

Papers

Total Citations

H-Index

About

Inamoto Kouya’s research sits at the intersection of robotics, computer vision, and machine learning, with a focused passion for enabling robots to autonomously understand and navigate their environments. His key contributions center on **unsupervised place discovery** for visual place classification, a critical challenge in robotic mapping and localization. Rather than relying on pre-labeled maps, Kouya’s work investigates how deep convolutional neural networks (DCNNs) can be used to intelligently partition a robot’s workspace into meaningful “places” without human intervention. His seminal paper on this topic (2016, 3 citations) and its follow-up (2017, 2 citations) directly address the open question of how to maximize classification performance—accuracy, precision, and recall—by optimizing the segmentation of space. While his citation counts are modest, his work tackles a foundational problem in autonomous navigation: enabling a robot to discover its own spatial categories. This research is particularly valuable for field robotics, where pre-mapping is impractical. Kouya’s contributions offer a principled, data-driven path toward more adaptive and self-sufficient robotic systems, making his work a thoughtful building block for students and researchers exploring unsupervised learning in robotics.