Meet Graava: The First Video Camera Using Artificial Intelligence to Identify and Automatically Edit …
A U.S. startup has developed a smart action camera that can automatically find the most exciting parts of a video and weave them together.
Graava uses sensors and software to detect the best scenes and can upload the edited version of the video to the cloud.
Aside from everyday action cam use, the device could be particularly useful for videos of road rage incidents like one recently recorded in England that has been viewed over 4 million times.
The camera links to a smartphone app through which users can specify the length of the edited version of a video. For instance, they can tell the app to summarize three hours of recordings into a 2-minute clip.
The camera’s algorithms will analyze the sensor data highlights, such as a sudden increase in recorded sound, to pick the best scenes from the footage. Users are notified when the edit is ready for sharing through social media.
The software isn’t perfect, of course, and can miss scenes that users think are interesting. Users can browse the recording and manually add the scenes they want with a drag and drop interface.
About the size of a pack of cigarettes, Graava is small and light enough to attach to a baseball cap. It’s equipped with an accelerometer, gyro sensor, GPS, Wi-Fi and Bluetooth modules, as well as a speaker and microphones. It can also take heart rate monitor data from smartwatches including Apple Watch.
The image sensor can shoot 1080p HD video at 30 frames per second (fps) or 720p resolution video at 60 fps. The former will drain the battery in about three hours. To recharge it, the camera can be placed on a wireless charging mat.
Optional accessories include wall-mounting brackets so the camera can be used as a home monitor.
The Graava camera is available for a limited time at US$249, with shipping scheduled for February 2016. An optional subscription service allows users to store videos in the cloud and automatically combine recordings from multiple cameras into one video clip.
Via: Google Alerts for AI