Consent Preferences

Casting and Compliance: How to Recruit, Screen, and Legally Capture Human Scan Subjects

The complete guide to diversity planning, pre-scanning protocols, and ethical best practices for AI-ready 3D human datasets.

Introduction

The success of any 3D human dataset begins long before a camera shutter clicks or a scanner starts capturing.
The human subjects you select, the care with which you prepare them, and the legal/ethical framework you put in place form the foundation on which your AI-ready dataset is built.

In this deep-dive guide, we’ll explore every step of this second stage in dataset creation—from casting diverse participants to on-set scanning protocols—providing a clear roadmap for creating high-quality, legally compliant, and bias-aware human datasets.

This post assumes you’ve already defined your dataset blueprint (covered in Part 1 of our series). Now, we move into executing that vision with real people.

Casting for Diversity and Dataset Quality

Why Casting Matters in AI

A 3D dataset is only as useful as it is diverse.
Models trained on narrow, homogeneous datasets struggle to generalize—causing failures in real-world environments.

Well-executed casting prevents algorithmic bias and ensures a wide spectrum of human variation enters your dataset.

Diversity Dimensions to Plan For

  1. Demographics:
  • Age brackets (children, teenagers, adults, seniors)
  • Ethnic groups and skin tones representative of global populations
  • Gender identity (male, female, non-binary, trans)
  1. Body Diversity:
  • BMI spectrum (underweight to obese)
  • Height ranges
  • Body proportions: limb-to-torso ratios, hand size, foot size
  • Physical conditions: presence of prosthetics or mobility aids
  1. Cultural Features:
  • Hairstyles and head coverings
  • Clothing types (for clothing-specific datasets)
  • Tattoos or body art
  1. Environment & Activity:
  • Models accustomed to specific motion types (sports, dance, work-related movements)

Pro tip: Build a “representation matrix” (subjects vs. diversity attributes) to ensure coverage as the dataset grows.

Recruitment Strategy

  • Work with casting agencies that specialize in diversity, or create open calls in multiple regions.
  • Pre-screen using a digital form that captures:
    • Demographics
    • Hair type/length
    • Facial hair
    • Tattoos/piercings
    • Willingness to sign a biometric consent form

 

Balance the final pool to match your dataset’s target demographic composition.

Screening for Physical Features That Impact Scanning

Once the casting shortlist is ready, technical screening ensures you know how each physical attribute will behave under scanning conditions.

Hair

  • Loose curly or frizzy hair can create gaps in 3D reconstruction.
  • Wet, gelled, or shiny hair introduces reflections.
  • Consider scanning with both hair tied back and natural states.

Beards and Facial Hair

  • Beards create occlusion zones around the jawline.
  • Long facial hair requires high-res texture capture and consistent grooming during sessions.

Body Hair

  • Fine body hair, especially on arms and legs, affects specular highlights in photogrammetry.
  • Document this and capture with neutral diffuse lighting.

Makeup

  • Heavy makeup alters color calibration of skin textures.
  • Glossy lipstick or eyeliner adds unwanted highlights.

Best practice: For baseline datasets, prefer natural or minimal makeup.

Accessories & Piercings

  • Jewelry, glasses, or metallic piercings introduce strong reflections.
  • Remove when possible, or plan specific captures for datasets focusing on accessories.

Nails

Long nails or reflective nail polish interfere with hand and finger scans.

Pre-Scan Preparation: Before Stepping on Set

Clothing Instructions:

  • For baseline scanning, fitted neutral clothing (or specific attire if dataset requires).
  • Avoid strong patterns, reflective materials, and loose garments that occlude body shapes.

Skin Preparation:

  • Avoid oil-based lotions just before scanning.
  • Remove reflective makeup and jewelry.

Briefing:

  • Provide clear instructions on session flow and what poses will be captured.

Address privacy and comfort (changing rooms, gender-sensitive staff, and pose expectations).

Legal and Ethical Framework

Biometric data, like 3D scans, is heavily regulated. Ignoring this exposes projects to risk.

Informed Consent Process

  • Provide a plain-language consent form explaining:
    • Purpose of the dataset
    • Intended use (AI research, commercial, internal)
    • Data storage period
    • Withdrawal rights

Regulatory Compliance

  • GDPR (Europe), CCPA (US), and other local data protection laws govern biometric data.
  • Ensure:
    • Data minimization (collect only necessary data)
    • Secure encrypted storage
    • Access control

Special Considerations:

  • Minors: Require parental/guardian consent

Sensitive Data: Anonymize or pseudonymize whenever possible

On-Set Scanning Protocols

Capturing Baseline Data

  • Measure height and weight with calibrated tools.
  • Take neutral front, side, and back reference images for verification.

Scanning Workflow

  • Ensure multi-camera synchronization.
  • For photogrammetry, use non-reflective backdrops and polarized lighting.
  • Start with neutral T-pose/A-pose then capture dynamic or dataset-specific poses.

Metadata Capture

Record for each subject:

  • Demographics
  • Hair/facial hair state
  • Clothing worn
  • Lighting setup
  • Session date and scanner version

This metadata becomes critical for filtering, balancing, and retraining models later.

Documentation and Version Control

  • Store signed consent forms securely (digital or physical).
  • Version scans and keep a clear audit trail.

Maintain links between scans, metadata, and legal documentation.

Key Takeaways

  • Casting is strategic: diversity and planning reduce dataset bias.
  • Screen physical features in advance to avoid scanning issues.
  • Legally compliant processes protect both participants and projects.

Metadata-rich documentation adds value and accountability.

Next in the Series

The next article will cover the technical side of building a photogrammetry -> Part 03: Inside the Studio

🤝 Ready to Plan With Experts?

We’ve built production-grade datasets for AI, gaming, digital fashion, and more—scanning thousands of humans with precision and care.

Whether you’re prototyping a research model or deploying at enterprise scale, we help you plan and execute every step of your 3D dataset pipeline.

Contact us to discuss your project and get a free consultation or sample scan set.

Author picture

We bring deep expertise and precision to the art of capturing real people in digital form. Whether you're creating lifelike characters for games and films, or training AI with high-fidelity human datasets, we guide you through every step—from casting and scanning to metadata structuring and delivery.

Our mission is to help you build better products and smarter models by turning physical humans into richly detailed digital assets—ready for any pipeline.

View All Posts

About Us

At Digital Reality Lab, we bring deep expertise and precision to the art of capturing real people in digital form. Whether you’re creating lifelike characters for games and films, or training AI with high-fidelity human datasets, we guide you through every step—from casting and scanning to metadata structuring and delivery.

Our mission is to help you build better products and smarter models by turning physical humans into richly detailed digital assets—ready for any pipeline.

Recent Posts

Author

I specialize in capturing reality and turning it into data – from photogrammetry rigs to digital human datasets for games, research, and AI. When not building pipelines, I’m exploring nature, climbing, and searching for the next big idea.