Loading...

Microsoft Launches Seeing AI App to Assist People with Low-Vision Abilities

TL;DR

  • Microsoft’s Seeing AI app aids people with low vision by reading text, identifying objects, and describing scenes using AI.
  • It accurately reads text, recognizes faces, identifies currency, and describes surroundings, making it easier for users.
  • However, it may struggle with complex objects and specialized text, showing room for improvement.

On Thursday, at its Ability Summit, Microsoft shared that Seeing AI is now available in an additional 14 languages(total 33 languages) and will be rolling out the latest generative AI models, including richer descriptions of images and chat capability for photos and documents. The innovative app harnesses Azure AI technologies to empower those with low-vision abilities, offering enhanced assistance in navigating daily life.

Introduction to the Seeing AI app

The Seeing AI app integrates cutting-edge Azure GPT-4 Turbo with Vision AI capabilities, ensuring responsive and accurate responses on any smartphone equipped with these features. Designed to cater to various needs, the app provides a range of functionalities.

The app enables users to have text read aloud to them, facilitating easier comprehension of written content. Shorter text can be instantly read in live camera mode, while longer passages are captured, converted to text, and subsequently vocalized.

Versatile features

Bar Code Reader: Users can scan product UPCs to obtain relevant information.

Scene Description: The app provides descriptive insights into the user’s surroundings, focusing on the nearest objects.

Person Identification: Through facial recognition technology, the app describes individuals and identifies their emotional expressions.

Currency Recognition: Users can easily identify various bills from 17 different countries.

Color Identification: The app discerns colors captured in live photos, aiding users in distinguishing visual elements.

Handwriting Conversion: Handwritten notes are seamlessly converted into text for improved accessibility.

Brightness Detection: A tone varying in pitch depending on brightness assists users in perceiving their environment.

While the app boasts impressive accuracy in many scenarios, it is essential to acknowledge its limitations. The app may struggle with more complex items like coin sorters despite correctly identifying common objects such as shaving razors. Similarly, while text recognition on receipts and packaging is generally reliable, the app may falter when confronted with mathematical or scientific texts.

Implications and prospects

The Seeing AI app marks a significant step forward in digital accessibility, demonstrating Microsoft’s commitment to leveraging AI for social good. By harnessing the power of Azure AI technologies, the app empowers individuals with low-vision abilities to navigate their surroundings more independently.

AI and machine learning advancements hold promising prospects for further enhancing the app’s capabilities and addressing its current limitations. Continuous refinement and iteration will be crucial in ensuring that Seeing AI remains at the forefront of digital accessibility solutions, catering to the evolving needs of its users.

Disclaimer. The information provided is not trading advice. Cryptopolitan.com holds no liability for any investments made based on the information provided on this page. We strongly recommend independent research and/or consultation with a qualified professional before making any investment decisions.

Share link:

James Kinoti

A crypto enthusiast, James finds pleasure in sharing knowledge on fintech, cryptocurrency as well as blockchain and frontier technologies. The latest innovations in the crypto industry, crypto gaming, AI, blockchain technology, and other technologies are his preoccupation. His mission: be on track with transformative applications in various industries.

Most read

Loading Most Read articles...

Stay on top of crypto news, get daily updates in your inbox

Related News

Victoria
Cryptopolitan
Subscribe to CryptoPolitan