Crowdsourcing Case Studies > Video Recordings for AI Training of an Eyetracking Tool

Training Data for Eye Square’s
Eyetracking Tool

Case Study – Video Recordings for AI Training of an Eyetracking Tool

Case Study – Video Recordings of Faces and Eye Movements as AI Training Data for an Eyetracking Tool

With the help of thousands of Clickworkers, Eye Square has conducted its largest data collection to date for the new SEAL Tool (Smartphone Eye tracking Algorithm Tool).
The recorded faces and eye movements from the Clickworkers served as AI training data to make the eye tracking tool even more precise.

Get in touch with us! +1 (212) 878-6686 +49 201 95971830
Case Study – Creation and Analysis of Speech Recordings as Training Data for a Speech Recognition Software

The Challenge

Eyetracking systems are only as good as the training data they are based on. The challenge lies in optimizing and training the system for the various eye movements and reactions of the users. Without considering human behavior and individual movement patterns, an optimal eyetracking system cannot be developed.

The eye movements of the users / the later market research panelists must be captured precisely and quickly to achieve an accurate system response. Insufficient recognition leads to inaccurate or even incorrect test results. This would significantly impair efficiency and user experience and reduce the reliability of the market research results.

To optimize the eyetracking tool for a wide range of individual eye movement patterns, it requires the recording of face videos from a large number of diverse individuals.

The Solution

Our Clickworkers from various countries and regions use a mobile app provided by Eye Square, which records their faces and eye movements. The app displays various points on the screen that differ in size, color, direction, and speed. Participants follow these points with their eyes while the app records their movements. These recordings serve as training data for the eyetracking tool.

With these data, the manufacturer trains the eyetracking tool and optimizes it for the individually different movement patterns of the users. This ensures that Eye Square’s market research clients receive reliable and accurate data without frustrations caused by faulty recognition.

clickworker is happy to support you in the creation or validation of AI training data.

 

Contact us now!
Case Study – Creation and Analysis of Speech Recordings as Training Data for a Speech Recognition Software

Project Data

  • Goal:
    The goal of this project was to collect face videos in frontal view on Android smartphones to train Eye Square’s eyetracking tool. This project aimed for a wide ethnic diversity without specific quotas within the target countries.

  • Clickworker Qualifications:
    Participants from the target regions
  • Number of Participants:
    Several thousand from many regions, including Germany, the US, Brazil, South Africa, India, China and the wider Indo-Pacific region.
  • Task:
    • Step 1: Acceptance of the Consent Form
    • Step 2: Entry of their Clickworker ID with redirection to the test app
    • Step 3: Placing/Stabilizing the smartphone
    • Step 4: Performing the tasks:
      Following the moving points on the app screen with their eyes.
Case Study – Creation and Analysis of Speech Recordings as Training Data for a Speech Recognition Software

 

Ethical AI through Fair Payment

Ethical AI – Fairly Paid Training Data

Eye Square and Clickworker have jointly developed the claim “Ethical AI – Fairly Paid Training Data” to emphasize our commitment to fair compensation for the participants.

Eye Square requested an additional 100% bonus for each Clickworker, which we successfully implemented and paid out. As a result, the Clickworkers’ remuneration for the tasks was doubled. This particularly fair payment ensures higher satisfaction and motivation among Clickworkers, leading to better data quality and faster project completion. This is another step towards ethical AI development and fair working conditions.

We say thank you to Eye Square and congratulations on the great eyetracking tool developed with fairly paid AI training data!

Case Study – Ethical AI - Fairly Paid Training Data
Case Study – Creation and Analysis of Speech Recordings as Training Data for a Speech Recognition Software

Advantages

  • Fast Project Execution:
    Efficient and prompt implementation through structured processes.
  • Access to a Diverse Participant Group:
    Access to a large and diverse group of participants worldwide.
  • Quality Assurance through Test Runs:
    Smooth process through initial test phases and continuous quality controls.
  • Scalable Throughput:
    Flexibility in project size and scope to meet different requirements.
  • Flexible Workforce:
    Adaptable and diverse workforce for various project needs.
  • Personal Project Support:
    Individual support and assistance throughout the entire project by our experienced project management team.
Linkedin Crunchbase Youtube Blog Newsletter