All articles tagged with #object recognition
The article reviews Rokid’s new AI-powered AR glasses, highlighting their impressive features such as real-time translation, teleprompter mode, object and scene recognition, and native photo/video capture, all powered by AI and a Qualcomm processor. The glasses are lightweight, versatile, and suitable for content creators and travelers, with a promising suite of AR tools and a competitive price of $499 on Kickstarter.
Researchers at Penn State developed a human-inspired AI training method that mimics infant learning using spatial data, resulting in a 14.99% performance boost in object recognition. This approach, tested in virtual environments, could advance AI systems for extreme environments or space exploration.
A study from MIT reveals that early exposure to limited color information helps the brain develop the ability to recognize objects based on luminance, which later allows for the incorporation of color without losing grayscale recognition. This adaptation explains why children who regain sight after congenital cataracts struggle with black-and-white object recognition due to an overreliance on color cues.
Google is reportedly working on an AI assistant called "Pixie" that will be exclusive to its Pixel devices, using data from Google products on the phone. It is expected to launch with the Pixel 9 and 9 Pro and could evolve into a more personalized version of the Google Assistant. Additionally, Google has discussed the possibility of developing glasses that utilize the AI's object recognition capabilities to provide guidance on various tasks.