Initially used to facilitate the accessibility of photographs to visually impaired users by means of their description, Facebook Lumos technology is now also applied to image searches in the social network through a new update of this function that for the time being is available to users in the United States both through the mobile application and through the web version.
The idea is that, through the use of Artificial Intelligence and Computer Vision, users can now search for the content available in the shared photographs themselves, beyond the titles and descriptions they have been carrying.
In this sense, users will be able to search for people with specific clothing, open places facing the sea, animal photographs, among many other possibilities, where little by little the type of activity being carried out by the people will be present in photographs. Those users will receive a diverse relationship of relevant photographs rather than photographs where there may be slight variations between them.
In the future Facebbok is planning to take Lumos to video search, but for now it has a long way to go to have enough training to be reliable enough and accurate enough to reach the general public.
To point out that Lumos is a technology developed for the compression of photographs and videos through Artificial Intelligence and Computer Vision. This technology already has a long journey in the identification of the available contents in millions of photographs, but still is needs to be further perfected. So, this technology will continue to receive improvements on a constant basis to be able to reach a more accurate level and provide the best result.