Contact Us

Augmented Reality

Crowdsourced data for gesture recognition improved model accuracy by 14%.

Industry and use case:

AR, gesture recognition

Data:

15,000 videos

Project duration:

21 days

Challenge:

The customer is developing a smart TV box with an RGB camera that recognizes users’ gestures. To improve the quality of gesture recognition, the company asked the Training Data team to collect and annotate 15,000 videos with 14 various gestures. The main challenge was to mine videos that mirror actual users’ gestures in real life when a user set up the TV box above the TV

Solution:

We crowdsourced video collection, while our team created detailed requirements for the video and organized strict quality control

Outcomes:

  • Gesture recognition accuracy was improved by 14%
  • We collected and annotated videos filmed by 7,000 different people
We came up with an idea, but we didn't have precise requirements. The CEO and CTO advised us and built the entire data collection process. The team independently developed the video collection requirements and crowdsourced the video collection. Very flexible team that allowed us to experiment and refine our needs to improve the final product!
Alex P.
Product Owner

Stages of work

  • Application

    /01
    Leave a request on the website for a free consultation with an expert. Th e acco unt manager will guide you on the services, timelines, and price
  • Free pilot

    /02
    We will conduct a test pilot project for you and provide a golden set, based on which we will determine the final technical requirements and approve project metrics
  • Agreement

    /03
    We prepare a contract and all necessary documentation upon the request of your accountants and lawyers
  • Workflow customization

    /04
    We form a pool of suitable tools and assign an experienced manager who will be in touch with you regarding all project details
  • Quality control

    /05
    Data uploads for verification are done iteratively, allowing your team to review and approve collected/annotated data
  • Post-payment

    /06
    You pay for the work after receiving the data in agreed quality and quantity

Timeline

  • 24 hours
    Application
  • 24 hours
    Consultation
  • 1 to 3 days
    Pilot
  • 1 to 5 days
    Conducting a pilot
  • 1 day to several years
    Carrying out work on the project
  • 1 to 5 days
    Quality control
You pay for the work after you have received the data
in the established quality and quantity

Why
Training Data

  • Quality Assurance:
  • Enhanced Data Accuracy
  • Consistency in Labels
  • Reliable Ground Truth
  • Mitigation of Annotation Biases
  • Cost and Time Efficiency
  • Data Security and Confidentiality:
  • GDPR Compliance
  • Non-disclosure agreement
  • Data Encryption
  • Multiple data storage options
  • Access Controls and Authentication
  • Expert Team:
  • 6 years in industry
  • 35 top project managers
  • 40+ languages
  • 100+ countries
  • 250k+ assessors
  • Flexible and Scalable Solutions:
  • 24/7 availability of customer service
  • 100% post payment
  • $550 minimum check
  • Variable Workload
  • Customized Solutions
woman

Tell us about your project!

    Choose interested services:

    Select an option

    • Data labeling

    • Data collection

    • Datasets

    • Human Moderation

    • Other (describe below)