Fei-Fei Li, chief scientist of artificial intelligence and machine learning at Google Cloud, announced the launch of a new machine learning API for automatically recognizing objects in videos and making them searchable at Google’s Next Cloud conference today in San Francisco.
the tool, which Google wants to make available to enterprises, would allow for videos to be searchable and discover able just like photos are currently on the Google Photo app. Naturally a user would search each shot, frame thanks to the tool, without relying manually, in order to find the exact video footage.In a demo,Google showed in a series of video how the tool could evaluate exact labels; when asked to find beach was able to find videos which had scenes from beaches in them, complete with time stamps. At least in Google’s demo, it was genuinely impressive. Google is making the API available to developers, just as it has with its other machine learning APIs.
At this point, it is an enterprise solution, a deep-learning tool built on frameworks like TensorFlow that companies can use to parse through their stored videos and extract metadata. If you wanted to hunt through your vast media collection to hunt down “tiger,” for example, you’d get a result like this :
Google also says the tool can detect scene changes within the video, and can help organisations with media archiving and boost content discovery for video. This API relies on Google’s current vision recognition models, which are also driving video search in YouTube.