AI Stats Generator

The AI Stats Generator (Coming Soon) analyzes match video and automatically detects volleyball plays — identifying players, tracking the ball, recognizing actions, and generating stat events. Think of it as an automated stat tracker that watches your game film.

How It Works

1. Upload a match video (or paste a YouTube URL)
2. Configure options — select which models to use, set roster info
3. Watch progress — the pipeline processes the video through multiple stages in real time
4. Review results — review detected plays, approve or reject each one, then import into your match stats

What It Detects

Detection	Description
Players	Tracks player positions throughout the video
Jersey Numbers	OCR reads jersey numbers to identify players
Ball	Tracks the volleyball for rally detection
Actions	Classifies attacks, serves, digs, blocks, and sets
Rallies	Detects rally boundaries (when play starts and stops)

Submitting a Job

From a File

1. Go to AI Stats in the Statistics submenu
2. Click Upload File and select your match video
3. Optionally set your roster (jersey numbers) for player identification
4. Click Submit

From YouTube

1. Switch to the YouTube URL tab
2. Paste a YouTube URL
3. Set a start time (MM:SS) and duration (default 10 seconds)
4. Click Submit

Tip

Start with short clips (10–30 seconds) to test the system before processing full matches. This helps you verify that the camera angle and video quality produce good results.

Reviewing Results

After processing completes, you’ll see:

Event Review

A list of all detected plays with:

• Player identification (jersey number)
• Action type (kill, dig, serve, etc.)
• Confidence score
• Timestamp in the video

For each event, you can:

• Approve — accept it as-is
• Edit — change the player, action, or outcome
• Reject — discard the detection

Annotated Video

An annotated version of your video showing:

• Player bounding boxes with jersey numbers
• Ball tracking
• Action labels when plays are detected
• Rally start/end boundaries

Court Heatmap

A visual heatmap showing where on the court actions were detected — useful for understanding positioning patterns.

Tips for Best Results

• Camera position matters — a high, center-court angle gives the best results
• 720p or higher — the AI needs enough resolution to read jersey numbers
• Steady footage — handheld shake reduces detection accuracy
• Include your roster — providing jersey numbers helps the AI match detected numbers to players
• Good lighting — indoor gym lighting is fine; avoid harsh shadows