AI-POWERED HANDBALL ANALYTICS

Turn Handball Broadcasts
into Tracking Data

Pivex extracts player positions, ball trajectory, and real-time analytics directly from existing broadcast video — no specialised cameras or sensors required.

SCROLL
0.0%

Tracking Accuracy

0ms

Inference Latency

0+

Player IDs per Frame

0+

FPS Support

THE PROBLEM

Handball is Flying Blind

While football and basketball teams make data-driven decisions every day, handball clubs are still navigating by instinct.

📹

Broadcasts Are Locked

Millions of hours of handball footage exist but yield zero machine-readable data without expensive manual annotation.

💸

Manual Tagging Costs Thousands

Human video analysts charge $500–$2,000 per match. At that rate, a full season of tracking data is simply unaffordable for most clubs.

🏢

Enterprise Systems Are Out of Reach

Hawk-Eye and STATS Perform charge $100,000+ per year and require stadium hardware installations — inaccessible to 95% of the market.

THE SOLUTION

Raw Video In. Structured Data Out.

Pivex runs your existing broadcast footage through a computer vision pipeline that automatically identifies every player and the ball — no manual labelling, no new cameras, no installation.

You get frame-level coordinates, velocity vectors, player IDs, and possession events exported as clean JSON or CSV — ready for your analytics dashboard, coaching tool, or research model.

📼
Broadcast Video
.mp4 · .mov · stream URL
Pivex AI Pipeline
YOLO · ByteTrack · Homography
📊
Tracking Data
JSON · CSV · WebSocket stream
LIVE DEMO

See Pivex in Action

Watch our AI pipeline extract full tracking data from a raw broadcast feed in real time.

HOW IT WORKS

Three Steps to Tracking Data

01

Upload Broadcast

Drop in any standard broadcast .mp4 or stream URL. No special cameras. No hardware. Works with existing footage.

02

AI Inference

Our YOLOv11 + ByteTrack pipeline detects and tracks every player and the ball, applying homography to map positions onto a 2D pitch.

03

Export Tracking Data

Receive structured JSON or CSV with frame-level coordinates, player IDs, velocities, and ball events via API or direct download.

CAPABILITIES

Everything You Need

A complete tracking data stack from broadcast video to structured analytics — in a single API call.

🎯

Player Tracking

Frame-accurate positions, velocities, and acceleration for every player on the court throughout the entire match.

Ball Tracking

Sub-pixel ball detection with trajectory prediction, shot detection, and possession attribution per player.

🗺️

2D Pitch Mapping

Homographic projection maps broadcast camera perspective onto a standardised top-down pitch coordinate system.

Real-time Mode

Live streaming inference at under 50ms latency for in-game analytics dashboards and coaching support.

📡

REST API & Export

Consume data as JSON or CSV, or stream via WebSocket. Integrates with any analytics tool or custom dashboard.

🆔

Jersey Recognition

Player re-identification across camera cuts using jersey number detection and deep appearance embeddings.

TECHNOLOGY

Built on Proven Computer Vision Foundations

We build on the same CV/ML stack used in world-class sports analytics — packaged into a single API call.

YOLOv11
Object Detection
ByteTrack
Multi-Object Tracking
OpenCV
Homography & Vision
PyTorch
Deep Learning Runtime
FastAPI
Inference API
ONNX Runtime
Model Optimisation
EARLY ACCESS

Get Tracking Data First

Join the waitlist. We're onboarding our first clubs and analytics teams now.