Brands have benefitted from artificial intelligence (AI) and machine learning (ML) as social media platforms and other engagement websites and apps study user preferences and tune in on their conversations in order to cater better services to them. Previously, AI and ML were only capable of reading still frames, however now, the algorithm can identify objects within videos which opens up new avenues for placement ads, accessibility features, etc. 

Google recently launched the AutoML Video Intelligence, a graphical interface that helps even ML novices train custom models easily to track and identify objects that are present in the video. After tweaking and bettering it for a few months now, Google has finally introduced these new advances on the YouTube platform. The feature allows tagging of products that are present in video clips and provide direct links to shop for those products.

This broadens the horizons for commercial websites which can now tag their products, provide direct shopping options, facilitating broader e-commerce opportunities in the app.

Joining the innovative party is Facebook which has recently announced their collaboration with Inria, the French National Institute for Research in Computer Science and Automation, with whom they have developed a new method called DINO. While DINO is still under construction, Facebook is already confident that the model is functional. "Our model can discover and segment objects in an image or a video with absolutely no supervision and without being given a segmentation-targeted objective," states the official announcement blog.

Facebook confessed that segmenting objects has been one of the greatest challenges faced by computer vision since to be able to correctly identify an object, it truly needs to know what it is. This is being accomplished by DINO by self-supervised learning and a suitable architecture making it a lot easier and uncomplicated.

Similar to YouTube's business model, Facebook's DINO will help Facebook display and tag products on their platforms with support from other features related to AR and visual tools that could lead to much more advanced, more immersive Facebook functions which will definitely help the company in incorporating further data and personalization.