Sports Science × Pose Estimation × XR

Teaching AI to see human movement.

From sports & exercise science to AI — I build markerless human pose estimation and AI-driven movement analysis, and break down the research here and on YouTube.

MediaPipe
OpenPose
ARKit
OpenSim
Unity
Meta Quest

Subscribe on YouTube Read the blog Contact

130+YouTube videos

10+international talks

EN · 日本語bilingual content

POSE DETECTED · 98.7%

Focus areas

🧍

Human Pose Estimation

MediaPipe, OpenPose, and the latest markerless methods.

Explore →🏃

Motion Capture & Biomechanics

Sports movement analysis and validation studies.

Explore →🤖

Computer Vision & AI

AI-driven movement analysis and modeling.

Explore →🥽

XR / Metaverse

Unity & Meta Quest development.

Explore →

Latest articles

The Finished Build: Body, Face, Fingers & Wrists From One Webcam — Avatar Talk2026年8月13日
The Avatar Talk capstone: body, face, fingers and wrists driving one VRM from a single webcam, all in the browser. Why an arm solve can’t roll the wrist, the thumb bending about the wrong axis, caching frames and counting drops, the neutral face where 0.3 is the new zero, the mirror rule — and an honest ~9 fps, not the 30 I’d claimed.
Face, Hands & Filtering: The Full Webcam Avatar — Avatar Talk Part 4 (MediaPipe → three.js)2026年8月13日
The Avatar Talk finale: pose, face and hands driving one VRM in a single browser render loop. The One Euro filter that beats the jitter-vs-lag trade, filtering landmarks at the source, MediaPipe’s 52 ARKit blendshapes mapped to VRM expressions, and finger bend-not-splay hand tracking.
Two Pose Pipelines, One Question — Validated, or Not?2026年8月5日
Same recipe — detect a pose, track it, classify it — but very different evidence. One 2026 paper grades a physical-therapy exercise against expert ground truth (RTMPose + GRU, ~96%); the other profiles a basketball move with no ground-truth labels, and honestly says so. What validates a pose pipeline isn’t the detector — it’s the ground truth.
Capturing Motion Isn’t the Finish Line — From Kinematics to a Modeled Injury Number2026年8月4日
You captured the motion — you’re not done. Two July 2026 papers trace the path from validated kinematics to a modeled injury number: an OpenCap markerless scoping review (51 studies), and a badminton-lunge ACL-load model (OpenSim + XGBoost, R² ≈ 0.91). Each inferential layer adds its own uncertainty.
Retargeting: Drive a VRM Avatar With Your Webcam — Avatar Talk Part 3 (MediaPipe → three.js)2026年8月3日
Part 3 of Avatar Talk: retarget MediaPipe pose landmarks onto a VRM avatar’s bone rotations, in the browser. Positions vs rotations, the MediaPipe→VRM coordinate flip and mirror, rest→target quaternions, world→local chains, frame-rate-independent easing, and the single-camera depth fix.
Loading a 3D VRM Avatar in the Browser — Avatar Talk Part 2 (three.js + three-vrm)2026年8月3日
Part 2 of Avatar Talk: load a real 3D VRM avatar in the browser with three.js and @pixiv/three-vrm. Why VRM’s standardized bone map matters, the loading cleanup that counts on a 15 MB model, two-line breathing, and the expression system for lip sync.

Featured video

Profile

I’m Takashi Fukushima, a researcher and developer working across sports & exercise science, motion capture, and computer vision. My focus is markerless human pose estimation and AI-driven movement analysis, alongside XR development with Unity and Meta Quest. I publish peer-reviewed research and break it down here and on my channel.

Human Pose Estimation (MediaPipe, OpenPose, latest methods)
Motion Capture & Sports Biomechanics
Computer Vision & AI movement analysis
XR / Metaverse development (Unity & Meta Quest)

Achievements

Positions & Affiliations

Specially Appointed Assistant Professor, Education Incubation Center, Tokyo Gakugei University
Executive Board Member, Asia Council on Exercise & Sports Science (ACESS)
Executive Board Member & Secretary, Global Metaverse Laboratory
CTO, Aletheia LLC
International Technological Partner, Sports Performance Analysis Association in Bharat (SPAAB) — Capacity Building Sessions; HALO (Human Analysis of Locomotor Optimization) Sessions 1／2／3

Commissioned & Funded Projects

FY2024 — Japan Sports Agency, Support Project for Reiwa Japanese-Style School Physical Education: enriching PE classes in the GIGA School environment (Theme 1) and inclusive PE classes for children with and without disabilities (Theme 2) — Secretary-General
FY2025 — MEXT commissioned project: Promoting the Use of Advanced Technology and Education Data for Next-Generation Schools (demonstration project on cutting-edge technology and education data) — Research Specialist (Engineer)
FY2025 — Support Project for Reiwa Japanese-Style School Physical Education (improving teaching methods for inclusive PE classes) — Deputy Research Director
FY2026 — MEXT commissioned project: Promotion of Moral Education that Fosters the Ability to Put One’s Way of Life into Practice — Specially Appointed Assistant Professor