Google AI helps to digitize old photos

Google AI помогает оцифровывать старинные фотографии

Google has recently launched in the practical activities of new machine learning algorithms for the proprietary artificial intelligence that controls the services of the Google Cloud Platform – and apparently, these solutions are already yielding positive results. Today the company released a progress report on a new version of their algorithms for AI, which are now involved in a special project with the newspaper The New Your Times – the project consists in digitizing old and not so old photos stored in a special underground vault, and which has many thousand copies since the mid-19th century and ending with those that were made a few months ago.

It is clear that the process of digitization itself is not something revolutionary and unusual – however, the singularity of the joint project of Google and The New York Times is that sorting, processing and placement of photos on a virtual platform deals exclusively with artificial intelligence Google Cloud Platform without human intervention. Moreover, as pointed out by the project staff Google artificial intelligence to cope with its task quite well, demonstrating a high degree of self-learning during sorting.

It is clear that this thought came to specialists of both parties in thinking about how to preserve and increase the collection of photos is because of accidents no one is immune and nothing. It should be noted that the current version of the learning algorithm represents the time between Cloud Visio API, therefore Google is actively promoting aspects of their services in this project.

Someone could see this advertising and the desire to promote their products further, however, the company can be understood in this quest – at least, The New York Times absolutely no problem with that. It remains to wait for the next update of the algorithms of self-learning artificial intelligence Cloud Platform and then we can form an opinion on the effectiveness and consistency of approach by Google to their work.

According to the materials: