Data Collection

Collecting Photos of Bald Men for a Medical Client

Image

We found a fast and efficient solution for gathering 200 sets of photos of bald men from different angles. Thanks to a detailed guideline, we annotated the images according to the Norwood scale with the highest accuracy.

Industry Medicine
Timeline 1.5 months
Data 200 annotated sets
Image
Industry Medicine
Timeline 1.5 months
Data 200 annotated sets

The Task

Imagine this: A medical company conducting advanced research on baldness approaches us with an unusual request.

They needed a dataset of bald men. Each set had to consist of five photos of the same person taken from different angles: front view, profile (both sides), top view, and back view.

But it wasn’t just about collecting photos—it was about creating a database where each image was “labeled” with medical precision. The reference point was the Norwood scale, a classification system for the stages of baldness. There was no room for error.

The Solution

  • 01

    Preparation Stage:

    • Studying the Norwood Scale: We dived deep into the topic, researching everything available about the scale.
    • Guideline Creation: With the help of medical experts, we developed an extremely detailed guide for our annotators, describing each stage of baldness down to the millimeter, complete with illustrations and diagrams.
  • 02

    Annotator Training and Preparation:

    • Every annotator had to pass an entry test on the Norwood scale (a 50-example test) before starting work.
    • We conducted hands-on training on annotating photos with different degrees of baldness, analyzing ambiguous cases, and discussing evaluation criteria.
    • After training, we calibrated the annotators’ work to ensure uniformity and precision.
    • Annotators were also trained in using spreadsheets, data processing platforms, and annotation tools.
    • To support the team, we set up a helpdesk for real-time expert assistance.
  • 03

    Data Collection:

    • A major crowd-platform served as the primary channel for collecting photos.
    • We created a clear and engaging task for participants, complete with detailed instructions and examples. To encourage high-quality submissions, we offered competitive compensation.
    • Our team closely monitored the process to promptly identify and address potential issues.
  • 04

    Data Validation:

    • We implemented a multi-layered verification system. First, an automated check ensured that all required photos were present and met technical specifications.
    • Then, experts manually annotated the images using the Norwood scale, cross-referencing them with our guidelines.
    • To guarantee accuracy, we conducted a cross-validation, where one set of experts reviewed the work of others.

The Result

  • Not only did we meet the deadline, but we also achieved near-perfect annotation accuracy. Thanks to thorough annotator training and a multi-step validation process, annotation errors were kept to a minimum—less than 1%.

  • The client approved all the data and was highly satisfied with the results.

Similar Cases

  • Image
    Image Annotation

    Optimizing Harvest Efficiency

    A client in the agriculture industry approached Unidata to automate the strawberry-picking process and increase farm efficiency. With cameras capturing […]

    Lean more
  • Image
    Audio Labeling services for ml Audio Transcription

    Speaker-Segmented Audio Annotation

    Automated transcription alone couldn’t handle the nuance of real conversations with background noise and interruptions. Our human-in-the-loop workflow ensured every […]

    Lean more
  • Image
    Audio Transcription

    High-Load Audio Transcription

    For one of our clients, we completed the transcription of 80 hours of audio files without using pre-labeling. The project […]

    Lean more
  • Image
    Audio Labeling services for ml Audio Transcription

    Banking Call Categorization

    To automate call categorization, one of Eastern Europe’s largest banks entrusted us with sensitive voice data covering credit, debit, deposits, and balances. We built a privacy-first annotation pipeline with in-house experts, multilayer validation, and weekly reporting to ensure both compliance and accuracy—enabling faster, smarter service automation.

    Lean more
  • Image
    Image Annotation

    Pose Estimation for Proctoring

    How do you teach AI to recognize when a student is cheating during an exam? By accurately annotating 6000 images of real exam scenarios — and that’s exactly what we did.

    Lean more

Ready to get started?

Tell us what you need — we’ll reply within 24h with a free estimate

    What service are you looking for? *
    What service are you looking for?
    Data Labeling
    Data Collection
    Ready-made Datasets
    Human Moderation
    Medicine
    Other (please describe below)
    What's your budget range? *
    What's your budget range?
    < $1,000
    $1,000 – $5,000
    $5,000 – $10,000
    $10,000 – $50,000
    $50,000+
    Not sure yet
    Where did you hear about Unidata? *
    Where did you hear about Unidata?
    Head of Client Success
    Andrew
    Head of Client Success

    — I'll guide you through every step, from your first
    message to full project delivery

    Thank you for your
    message

    It has been successfully sent!

    This website uses cookies to enhance your experience, analyze traffic, and deliver personalized content and ads. By clicking "Accept", you consent to the use of cookies, as described in our Cookie Policy. Please choose your cookie preference.