Media files matching approach with the help of computer vision/machine learning to discover like-minded people for social interactions

The present invention provides media files matching approach with the help of computer vision/machine learning to discover like-minded people for social interactions. Our new suggested approach is to use media files taken from a person's camera (from mobile phones) and discover media taken by others. The other media matched with the original one will have common characteristics which will allow them to interact socially and start a new warm conversation because they have something in common (a common 'ice-breaker' problem). Media is being analysed and compared based on generated meta-data tags and extra weights like created time relevance, basic search preferences and location. This approach has been practically implemented in the platform called 'Yupi' (a mobile app that can be downloaded from the App Store in Australia). The approach is based on pre-trained models (Machine Learning) using Supervised Learning and classification models, and, specifically, Multiclass Classification. Therefore, use the above as computer vision to extract objects presented in media files. THE DRAWINGS Moments from Compass icon different users Resulting selection is being prepared on the background per each moment separately A Moment Correlating with'Water'on it Moments Figure 1 is showing how a user finds similar multimedia 'Moments' using the friending /social interaction platform called 'Yupi' (https://yupi.com.au/).