Adrenaline Icon

AI Vision Assist Scoring

Early Access Feature

AI Vision Assist Scoring is currently in development and not yet fully available to the public. This documentation describes the planned functionality for this upcoming feature.

Adrenaline's AI Vision Assist Scoring is a proprietary photo-finish tool that uses computer vision to help race officials make accurate, defensible calls when riders cross the finish line simultaneously. The AI doesn't make decisions—it provides visual evidence and suggestions to assist human officials.

The Philosophy: Assist, Never Replace

Adrenaline's AI philosophy is simple: extend attention, not judgment. The AI's job is to slow the moment down and surface evidence so officials can make better decisions. The human official always has the final say.

"No one can manually score 16 riders crossing simultaneously. What you really need in that situation is slo-mo to give you time to deduce what actually happened. That cannot happen without integrated video and it cannot happen without our proprietary AI Vision Photo-Finish Tool."

— From the Adrenaline About Page

How It Works

When riders cross the finish line, the AI Vision system analyzes the integrated video feed to detect vehicles and attempt to identify number plates. The system then presents this information to the race official for review and confirmation.

For Race Officials

Officials see an animation of the machine crossing along with the AI's prediction for the number plate. This gives them:

  • Slo-mo playback: Frame-level control to review close finishes
  • AI predictions: Suggested number plate identifications with confidence levels
  • Visual evidence: Video backing for every crossing
  • One-tap corrections: Tap to assign, confirm, or correct any crossing
  • Timestamped audit trail: Every decision is logged for accountability

For Spectators

Spectators see a camera icon next to each crossing. Clicking it shows what the AI vision system saw for that crossing, providing transparency into how the finish order was determined.

Key Features

Integrated Video + AI

The system requires fully integrated video. Unlike standalone timing systems, Adrenaline's AI works directly with the video feed to provide frame-by-frame analysis.

Number Plate Recognition

In ideal conditions, the AI attempts to read number plates and suggests identifications. Officials can accept, reject, or modify these suggestions with a single tap.

Adaptive Detection

In poor conditions (dust, mud, rain), the system gracefully downgrades to strictly detect the presence of vehicles without attempting number plate recognition. This still provides valuable slo-mo evidence for officials.

Anomaly Highlights

The system highlights unusual crossings (very close finishes, potential misreads) for quick official review.

Visual Receipts

Every crossing has a corresponding video frame. Officials can defend any call with visual evidence and spectators can review what the system saw.

Timestamped Audit Trail

Every edit, confirmation, or correction is logged with a timestamp. If a call is disputed, the complete history is available for review.

User Interface

Officials View

In the race officiate page, crossings are displayed in a list. Each crossing shows:

  • Lap number and crossing time
  • Assigned rider number
  • Camera icon to view AI vision analysis
  • Confidence indicator (when AI makes a prediction)

Clicking the camera icon opens the AI vision viewer, showing:

  • Video animation of the crossing
  • AI's predicted number with confidence percentage
  • Alternative predictions (if confidence is low)
  • Frame controls to step through the crossing frame-by-frame
  • One-tap number assignment or correction

Spectator View

On the public race spectator page, crossings show:

  • Lap number and crossing time
  • Confirmed rider number
  • Camera icon indicating AI vision is available

Clicking the camera icon shows what the AI vision saw (read-only view for transparency).

Handling Difficult Conditions

The AI Vision system is designed to adapt to real-world race conditions:

Ideal Conditions

Clear visibility, good lighting, unobstructed view of number plates. The AI confidently identifies numbers and provides high-confidence predictions.

Moderate Conditions

Some dust, mud splash, or partial obstructions. The AI provides lower-confidence predictions or multiple possibilities. Officials review and confirm manually.

Poor Conditions

Heavy dust, mud, rain, or low visibility. The system downgrades to vehicle detection only—it doesn't attempt number recognition but still provides slo-mo video evidence of crossings for manual review.

"What if it's too dusty? Too muddy? Or too much rain? It knows. Under the hood, the computer vision can downgrade features to strictly detect the presence of a vehicle and forgo number-plate recognition. That level of detection is still enough to assist the race official in isolating and analyzing all crossings and provide the evidence to back it up. It's the slo-mo that counts; when analysis isn't possible in real-time with human eyes, get a little help from your friend."

— From the Adrenaline About Page

Benefits for Officials

  • Confidence in close calls: Video evidence backs every decision you make.
  • Faster processing: AI suggestions speed up routine crossings, letting you focus on the difficult ones.
  • Reduced disputes: Visual receipts provide indisputable evidence for contested calls.
  • Less stress: The pack bunching up is no longer a nightmare—slo-mo gives you time to think.
  • Audit trail: Timestamped records protect you if a call is questioned later.
  • Glove-friendly controls: UI designed for quick taps, not precise clicking.

Requirements

Integrated Video System

AI Vision requires Adrenaline's fully integrated video capabilities. The system needs a dedicated camera angle capturing the finish line to analyze crossings in real-time.

Finish Line Camera Placement

The finish line camera must have an unobstructed view of crossing vehicles with number plates clearly visible (when conditions permit).

Processing Power

The AI Vision system requires sufficient local processing power to analyze video frames in near real-time. Specific hardware requirements will be documented upon release.

What AI Vision Is NOT

  • NOT automatic scoring: The AI doesn't make final decisions. Officials review and confirm every crossing.
  • NOT a replacement for officials: Human judgment is always required. The AI assists, it doesn't officiate.
  • NOT 100% accurate: Like any AI system, it can make mistakes. That's why officials review everything.
  • NOT a magic solution: In extremely poor conditions (heavy rain, complete darkness), even vehicle detection may not be possible.
  • NOT standalone: This feature requires Adrenaline's integrated video system—it won't work with external cameras or third-party video.

Common Questions

When will AI Vision be available?

AI Vision Assist Scoring is currently in development. There is no confirmed release date at this time. Check the Adrenaline website or contact support for updates.

Does this work with my existing cameras?

No. AI Vision requires Adrenaline's fully integrated video system. External cameras or third-party video feeds are not supported.

Will there be an additional cost?

Pricing for AI Vision has not been announced. It may be included in certain subscription tiers or available as an add-on. Details will be provided closer to release.

Can the AI make mistakes?

Yes. Like all AI systems, it can misread numbers or miss detections—especially in poor conditions. That's exactly why human officials review and confirm everything. The AI is an assistant, not a replacement.

What happens if I disagree with the AI's prediction?

You're the official—you have the final say. Simply tap to assign the correct number, and the system logs your decision. The AI's prediction is just a suggestion.

Does this require internet connectivity?

Like all Adrenaline race operations features, AI Vision is designed to work offline. Processing happens locally on your device. Internet is only needed to sync results after the race.

Can spectators see the AI predictions?

Spectators see what the AI vision captured (the video frame), but AI predictions and confidence scores are only visible to officials. This prevents confusion and maintains official authority over scoring decisions.

Learn More

To learn more about Adrenaline's philosophy on AI and how we're building assistance tools (not automation), visit:

Tags
raceaivisionscoring

Was this information useful?

Help us improve our documentation by rating this article and sharing your feedback.