As Machine Learning is making inroads into applications across both consumer and enterprise domains, Google is pushing updates to its Machine Learning APIs via its advanced models that help analyze and make sense of images and videos.
Cloud Video Intelligence API addresses the difficult problem of extracting entities from your video files. It extracts out the entities along with specific moments in the video that they were found. This helps create applications that can now search and index video content in an easy way. The API which was available in Alpha has now been promoted to Beta along with improvements in label detection along with identification of adult content. It also sees pricing updates with label and adult content detection free with first 1000 minutes. Shot Detection is also free for first 1000 minutes. Beyond that, a tiered pricing model applies. Check out the test drive link for trying out the Video Intelligence API.
The Cloud Vision API that has been around for a while provides features like label, logo detection, OCR, Face detection, Landmark detection via a simple to use REST API. The updates see significant improvements in accuracy for the specific features along with better response times. The new features released include web and document detection, that allows you to search for similar documents on the web, based on the metadata. The features will be rolled out gradually across the coming weeks to all users.
The Cloud Machine Learning APIs from Google and similar APIs from other vendors is a great way for organizers to start implementing Machine Learning in their applications. With sophisticated models exposed via REST APIs, it presents a less difficult learning curve while allowing the organizations to focus on their business use case. We do expect that these APIs will continue to benefit in terms of accuracy and response times, as the data set keeps getting large and with constant feedback from the early adopters.