News

Our paper BFMD is now available on arXiv.

Joined Nagoya Institute of Technology as Assistant Professor.

About Me

Ning Ding
Ning Ding

Education & Career

  • Assistant Professor, Graduate School of Engineering, Nagoya Institute of Technology
  • Ph.D., Graduate School of Informatics, Nagoya University
  • M.S., School of Information and Production Systems, Waseda University
  • B.S., Guangdong University of Technology

Research

Player performance evaluation and match video understanding in racket sports.

✉ Email GitHub Google Scholar CV / Resume

Research Overview

Sports Video Understanding, Dataset Construction, Player Evaluation, and Automatic Commentary

🎯 Tracking🏃 Pose↗ Trajectory ⚙ Tactics🧠 Reasoning💬 Commentary 🏆 Benchmark📊 Analysis
1Dataset Construction
Raw Match Videos
▶
Player ID
Shuttle / Ball
Court / Pitch
Skeleton Pose
Structured Annotations
▶
Serve Rally Smash Drop Clear Net Play
Event Segmentation
▶
  • ▶ Video Clips
  • ⬜ Tracking Boxes
  • 🦴 Pose Skeletons
  • ↗ Trajectories
  • 🏷 Event Labels
  • 💬 Captions
  • 📋 Metadata
Multimodal Dataset
19Matches
4Tournaments
8K+Shot Events
5Modalities
🧍 Pose↗ Trajectory🎯 TrackingRich · Scalable · Aligned

My
Research

📊Data
🤖Models
📈Evaluation
💬Language
2Player / Match Evaluation

Player Comparison

Attack
86 / 82
Defense
72 / 79
Consistency
88 / 81
Stamina
75 / 85
Clutch
90 / 84
■ Player A■ Player B

Performance Radar

Attack Defense Stamina Clutch Tactics

Activity Heatmap

Tactical Map

Form Over Time (Performance Score)

Match 1 Match 20
■ Player A■ Player B
3Captioning / Commentary

🏸 Badminton Commentary

"A sharp cross-court drop puts the opponent under pressure."

Confidence: 0.92

⚽

⚽ Soccer Commentary

"The striker breaks through the defensive line and finishes clinically into the bottom corner."

Confidence: 0.94

Shot Type Rally Phase Player Action Tactical Intent
👁Video Understanding
(Perception)
→
🧠Reasoning
(Context & Intent)
→
💬Language Generation
(Commentary)
4Multi-Sport Understanding

Badminton

🏸 Smash⬇ Drop↩ Clear🔗 Net Play🔄 Drive

Knowledge Transfer
& Generalization

⟵⟶

Shared Representations
Stronger Models

Soccer

↗ Pass⚽ Dribble✕ Cross🥅 Shot🛡 Defense
2Sports
19+Matches
10+Action Types
↑Cross-Sport Transfer

Toward broader sports understanding

🎯 Goal Build data-centric, intelligent systems that understand sports like humans.
📱 Applications Highlight Generation · Performance Analysis · Coaching Support · Broadcast & Media · Sports AI Agents
📈 Impact Advancing AI for sports and enriching the fan experience.

© 2025 Ning Ding · Nagoya Institute of Technology