Hiroshi Murase

About

Hiroshi Murase is a prominent researcher in computer vision and autonomous robotics, whose work spans visual recognition, robot localization, and human-robot interaction. His most influential contribution, "Subspace Methods for Robot Vision" (1996, 206 citations), revolutionized the field by reframing visual recognition as an appearance-matching problem rather than shape-based analysis — a paradigm shift that proved foundational for subsequent machine vision research. Building on this, his early work on learning algorithms for vision-based robot positioning established robust frameworks for mapping appearance to spatial location, earning nearly 100 additional citations across related publications. Murase's research evolved to address the practical challenges of autonomous navigation, with notable contributions in hybrid localization techniques that fuse probabilistic Monte Carlo methods with deep learning-based end-to-end approaches. His more recent investigations extend into reliability estimation for vehicle localization using convolutional neural networks and data-driven navigation of personal mobility vehicles in complex pedestrian environments. Complementing his engineering focus, Murase has also explored the human dimensions of autonomous systems, examining pedestrian gaze behavior during automated vehicle interactions and how driver experience shapes eye-gaze patterns in robotic wheelchair operation. Across three decades, his interdisciplinary contributions have meaningfully advanced both the theoretical foundations and real-world deployment of intelligent autonomous systems.

Research Focus

Key Achievements

Top Papers

1. 1
   Subspace methods for robot vision

   206 citations · 1996
2. 2
   Learning, positioning, and tracking visual appearance

   99 citations · 2002
3. 3
   Reliability Estimation of Vehicle Localization Result

   30 citations · 2018
4. 4
   Mobile Robot Localization Considering Class of Sensor Observations

   25 citations · 2018
5. 5
   Hybrid Localization using Model- and Learning-Based Methods: Fusion of Monte Carlo and E2E Localizations via Importance Sampling

   25 citations · 2020
6. 6
   Dynamic Active Search for quick object detection with pan-tilt-zoom camera

   15 citations · 2002
7. 7
   Personal Mobility Vehicle Autonomous Navigation Through Pedestrian Flow: A Data Driven Approach for Parameter Extraction

   12 citations · 2018
8. 8
   <title>General learning algorithm for robot vision</title>

   12 citations · 1994
9. 9
   What Is the Gaze Behavior of Pedestrians in Interactions with an Automated Vehicle When They Do Not Understand Its Intentions?

   8 citations · 2020
10. 10
    An Analysis of How Driver Experience Affects Eye-Gaze Behavior for Robotic Wheelchair Operation

    7 citations · 2019

Key Collaborators

Contact & Links