Mindtech’s Digital Humans 2.0 significantly reduce bias in AI Computer Vision Systems

27 Oct 2023

Share article:


At Mindtech, we’re taking AI computer vision to the next level with the release of our Digital Humans 2.0 technology.

To ensure AI systems understand and identify human behaviours accurately, diverse training data is essential. Our Digital Humans 2.0 offering is part of our Data Ops Platform and represents a significant leap forward in addressing bias and improving AI system accuracy.

Our Digital Humans achieve this diversity by varying elements of what we call “Digital DNA,” including skin tone, ethnicity, age, physical size, and clothing. We use the Monk skin tone scale to measure diversity, ensuring that our training data is an accurate representation of the world’s population. This level of diversity surpasses what can be achieved with real-world data alone and allows for an unlimited combination of variations.

Digital DNA vs Gen AI Models

Digital DNA technology outshines traditional “avatars” created by Generative AI models. 

These avatars often suffer from biases, rights issues, and visual artefacts that hamper model performance. Mindtech’s Digital Humans are precise and unbiased, offering controlled diversity in training data.

Practical Applications

Mindtech’s Digital Humans 2.0 will advance AI computer vision by reducing bias and enhancing diversity. As AI technology advances, addressing bias is crucial, and our innovative approach is paving the way for more accurate and fair AI systems.

Our Digital Humans have a wide range of applications across industries, including:

  • In-Cabin Automotive Testing: Our technology ensures that automotive AI systems are inclusive and unbiased by testing a range of variations in heights, faces, and behaviours.
  • Retail Spaces: In retail, especially in “just walk out” stores, Digital Humans guarantee that algorithms are fair and inclusive for all customers.
  • Lost Child Identification: In crowded places like shopping malls, our Digital Humans improve the identification of lost children, enhancing security and safety.

The ability to create diverse digital humans in precise and controlled ways is revolutionary for the creation of data sets,” commented Chris Longstaff, VP of Product at Mindtech. “Our platform, analyses and identifies bias in existing datasets, and allows data scientists to rapidly address diverse digital humans, improving AI model accuracy.

AI Expo Tokyo

The Mindtech team is attending AI Expo Tokyo, Japan’s premier trade show for AI tech companies this week. We are committed to advancing global AI technology and we’re looking forward to being part of NexTech Week Tokyo, a comprehensive tech event featuring AI, blockchain, and quantum computing.

The future of AI is brighter with Mindtech, leading the charge toward inclusivity and accuracy in the AI world. Contact us about our Digital Humans offering here and learn more about the AI Expo Toyko here.