PeopleofSuperintelligence

Researcher

Ryo Hachiuma

Ryo Hachiuma is a computer vision researcher whose work spans real-time SLAM, action recognition, and multimodal models for images, video, and autonomous driving. He coauthored DetectFusion and earned a Ph.D. in Engineering at Keio University under Hideo Saito; he also coauthored a method for estimating 3D human-object interactions from a wearable camera. His papers address image-caption bias, region-level image and video understanding, and cooperative driving with multimodal language models. As of October 2026, he is a Senior Research Scientist at NVIDIA Research Taiwan; his interests include video understanding and vision-language models, and five of his papers were accepted at CVPR 2026.

Now at
NVIDIA
Works
97
Citations
666
h-index
13
Badges

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2019

    DetectFusion at BMVC

    In 2019, he coauthors DetectFusion with Christian Pirchheim, Dieter Schmalstieg and Hideo Saito, a real-time RGB-D SLAM system for semantically known and unknown moving objects. BMVC selects the paper for its 2019 spotlight session.

  2. 2019

    An APMAR Best Paper Award

    He wins the Best Paper Award at the Asia Pacific Workshop on Mixed and Augmented Reality in 2019.

  3. 2020

    A MICCAI foot-ulcer win

    In 2020, he takes first place in MICCAI’s Diabetic Foot Ulcer Challenge. He coauthors Deep Learning in Diabetic Foot Ulcers Detection: A Comprehensive Evaluation, which compares the winning teams’ methods, including Faster R-CNN variants, YOLOv5 and EfficientDet.

  4. 2021

    A Keio Ph.D. and pose policy

    He earns a Ph.D. in Engineering at Keio University under Hideo Saito in 2021 and also works as a visiting researcher at TU Graz and Carnegie Mellon University. At NeurIPS 2021, he coauthors Dynamics-Regulated Kinematic Policy with Zhengyi Luo, Ye Yuan and Kris Kitani, a method for estimating physically plausible 3D human-object interactions from a single wearable camera.

  5. 2023

    Action recognition at Konica Minolta

    Before joining NVIDIA Research, he works as a computer vision engineer at Konica Minolta in Japan on human action recognition. At CVPR 2023, he coauthors the paper Unified Keypoint-Based Action Recognition Framework via Structured Keypoint Pooling with Fumiaki Sato and Taiki Sekii. The framework treats time-series human skeletons and nonhuman object contours as an input 3D point cloud through a tracking-free architecture.

  6. 2024

    NVIDIA Research Taiwan

    NVIDIA Research Taiwan is the APAC research team founded in 2022, where he works as a Research Scientist on multi-modal AI. His interests include video understanding and vision-language models. At EMNLP 2024, he coauthors From Descriptive Richness to Bias, which finds that generative image-caption enrichment increases gender bias and hallucination.

  7. 2025

    Omni-RGPT and V2V-LLM

    In 2025, he coauthors Omni-RGPT with Miran Heo, Min-Hung Chen, Yu-Chiang Frank Wang and others, a multimodal large language model for region-level understanding of images and videos. He is the second author of V2V-LLM, which fuses perception information from multiple connected autonomous vehicles in a multimodal large language model to answer driving questions about grounding, notable-object identification and planning.

  8. As of October 2026

    Senior Research Scientist at NVIDIA

    As of October 2026, he is a Senior Research Scientist at NVIDIA Research Taiwan, working on multi-modal AI. CVPR 2026 accepts five of his papers.

Career

  1. Since 2024NVIDIASenior Research Scientist
  2. 2023Konica MinoltaComputer Vision Engineer
  3. 2021Graz University of TechnologyVisiting Researcher
  4. 2021Carnegie Mellon UniversityVisiting Researcher

Education

  1. 2021Keio UniversityPh.D. in Engineering

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

Most-cited works

  1. Deep Learning in Diabetic Foot Ulcers Detection: A Comprehensive Evaluation2020 · 174 citationsWith Eibe Frank
  2. Unified Keypoint-Based Action Recognition Framework via Structured Keypoint Pooling2023 · IEEE Computer Society Conference on Computer Vision and Pattern Recognition · 59 citations
  3. Prompt-Guided Zero-Shot Anomaly Action Recognition using Pretrained Deep Skeleton Features2023 · IEEE Computer Society Conference on Computer Vision and Pattern Recognition · 48 citations
  4. Dynamics-Regulated Kinematic Policy for Egocentric Pose Estimation2021 · 44 citations
  5. Toward Unsupervised 3d Point Cloud Anomaly Detection Using Variational Autoencoder2021 · International Conference on Image Processing · 26 citations

From OpenAlex, as of 21 October 2026.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Ryo Hachiuma. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  2. Google has the highest combined h-index among AI organizationsIncludes NVIDIA. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  3. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes NVIDIA. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.
  4. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes NVIDIA. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Ryo Hachiuma yet. Send us a free-licensed photo.

  1. DetectFusion: Detecting and Segmenting Both Known and Unknown Dynamic Objects in Real-time SLAM arxiv.org. Checked 21 October 2026.
  2. Ryo Hachiuma ryohachiuma.github.io. Checked 21 October 2026.
  3. Deep Learning in Diabetic Foot Ulcers Detection: A Comprehensive Evaluation arxiv.org. Checked 21 October 2026.
  4. Dynamics-Regulated Kinematic Policy for Egocentric Pose Estimation arxiv.org. Checked 21 October 2026.
  5. Ryo Hachiuma | NVIDIA Research Taiwan research.nvidia.com. Checked 21 October 2026.
  6. Unified Keypoint-Based Action Recognition Framework via Structured Keypoint Pooling (CVPR 2023 Open Access Repository) openaccess.thecvf.com. Checked 21 October 2026.
  7. NVIDIA Research Taiwan research.nvidia.com. Checked 21 October 2026.
  8. From Descriptive Richness to Bias: Unveiling the Dark Side of Generative Image Caption Enrichment | NVIDIA Research Taiwan research.nvidia.com. Checked 21 October 2026.
  9. Omni-RGPT: Unifying Image and Video Region-level Understanding via Token Marks arxiv.org. Checked 21 October 2026.
  10. V2V-LLM: Vehicle-to-Vehicle Cooperative Autonomous Driving with Multimodal Large Language Models arxiv.org. Checked 21 October 2026.