Google announced that the beta phase of its Cloud Video Intelligence API will now be open to everyone.
The Video Intelligence API does can do two things: determine the shots (scene changes) in a video, and assign labels to the video and individual shots. Together with putting the API in public beta phase, Google also added support for detecting adult content.
Labeling and detecting adult content costs 10 cents per minute of video, and shot detection costs 5 cents per minute of video. Google Cloud platform users get both services for free for the first 1000 minutes of video they upload. To analyze a video users have to put it in the Google Cloud Storage, and the API supports common video formats such as .MOV, .MPEG4, .MP4, and .AVI.
Google also announced some changes to its Cloud Vision API, which operates on still images. The label detection model was extended and now contains over 10.000 different labels. With an update of the "safe search" model, adult content errors were reduced by 30%. Text detection has a 5% improvement and processes 25% faster. Emotion detection accuracy has improved for the emotions sadness, surprise, and anger. Although these updates were announced at the end of June, updates will still take three weeks to roll out globally.
For an example of the Video Intelligence API, more information, and how to get started, go to the Google Cloud platform. The Google Cloud platform also contains all information about the Cloud Vision API.