smart glasses Probably failed to take offbut add Artificial intelligence (AI) It could be the key to developing truly innovative wearable technology.
In the U.S. and Canada, Ray-Ban Meta smart glasses now feature multimodal AI technology with software called “Meta AI Virtual Assistant.” By multimodal AI, we mean a device with generative AI that can process queries involving multiple media (both audio and images). It allows the wearer to respond appropriately to queries based on what they see.
“Say you're traveling and trying to read a menu in French. Smart glasses use a built-in camera and Meta AI to translate the text and give you the information you need without having to pull out your phone or stare. . on screen,” a representative for Meta explained in an April 23 statement.
Related: Smart glasses could enhance privacy by replacing cameras with this 100-year-old technology
The device first takes a photo of what the wearer is looking at, and then AI uses cloud-based processing to tell the wearer a voice, such as, “What type of plant am I looking at?” Provide answers to questions.
Meta first explored integrating multimodal AI into its Ray-Ban Meta smart glasses, which were released in limited quantities in 2016. December 2023.
Our reporter is testing the device's AI capabilities. The Verge I found that when I asked for the car model, it answered correctly most of the time. For example, you can explain the types of cats and their characteristics with images taken with a camera. However, the AI had trouble accurately identifying the types of plants one reporter owned, and struggled to accurately identify a groundhog in his neighbor's backyard.
multifaceted conspiracy
AI-powered virtual assistants are nothing new, with Google Assistant, Amazon Alexa, and Apple's Siri all providing smart answers to queries in natural language. But the heart of Ray-Ban smart glasses' Meta AI is their multimodal capabilities.
The ability to fuse and process data from multiple sensor modules (such as cameras and microphones) Multimodal AI can produce more accurate and sophisticated results Versus unimodal AI systems. Google Gemini Multimodal AI Model For example, process some cookie photos, Reply with recipe.
Multimodal AI is trained to identify patterns in different types of data input through multiple neutral networks (a collection of machine learning algorithms arranged to mimic the human brain), including text, images, audio, etc. It can process input data from.
In smart glasses, the combination of sensors on the glasses and these neural networks means that AI can understand the world the wearer sees. As a result, the system can respond to more sophisticated queries and provide smarter contextual information.
However, in the case of Ray-Ban Meta devices, there is still some distance before the AI can match the AI processing power found in modern smartphones. These benefit from a more powerful chipset and onboard sensor fusion. Data is acquired from multiple sensors and processed together. For example, camera apps can provide scene recognition to intelligently adjust lighting and color balance, or combine data from thermometers and optical sensors. Give you better feedback about your workouts with your smartwatch.