Last week, Google announced that it has introduced new enhancements for its Vision AI portfolio, comprised of AutoML Vision Edge, AutoML Video, and Video Intelligence API. Together, these three services from Google provide machine learning (ML) solutions for a wide variety of businesses. Google’s Video Intelligence API provides ML models that can automatically recognise objects, scenes, and actions from videos. Now, Google has added a feature to Video Intelligence API that can detect, recognise, and track more than 100,000 logos.
AutoML Video is a more complex platform that allows businesses to train custom models and make their own labels for identifying video content. Google’s enhancement now allows AutoML Video to detect and track multiple objects between frames. Google considers this to be a particularly important component for uses of AutoML Video in applications like traffic management and robotic navigation.
Besides adding features to the already extant programs for videos and ML, Google has also promised to roll out AutoML Vision Edge, a program that businesses can use to create and deploy ML models “at the edge.” While still in beta testing, AutoML Vision Edge can classify images and, now, detect objects as well. “Object detection is critical for use cases such as identifying pieces of an outfit in a shopping app, detecting defects on a fast-moving conveyor belt, or assessing inventory on a retail shelf.” AutoML Vision Edge boasts low latency and high accuracy with a small memory footprint, making it prime for integration within augmented reality (AR) applications.
TRYON, a start-up that specializes in AR software for e-commerce and retail stores, uses AutoML Vision Edge to create AR shopping experiences for customers who want to try on jewelry in the comfort of their own homes before making a purchase.
“We evaluated AutoML Vision Edge Object detection and were so impressed with the accuracy and the speed that we decided to include the object detection model in our first beta release.,” said Andrii Tsok, co-founder of TRYON. “By integrating AutoML Vision Edge Object detection into our platform we expect to double our productivity by reducing the amount of resources and time for managing internal infrastructure.”
Yet to be released to the general public, the addition of AutoML Vision Edge to Google’s Vision portfolio promises a versatile and powerful toolkit for AR applications.
