One of our clients developed a web platform where people could find their old classmates and contact details just by providing a photo of them. The project needed an object recognition software that would select all the pictures from Yearbook Albums in order to eventually match them to the input picture of the user.
The recognition software based on artificial neural network models is capable of self-optimizing pattern recognition processes while going through more and more data.
The software that we developed was capable of recognizing different geometrical shapes from the albums, whether they were square, round or other polygon shapes, using classic image restoration algorithms to sharpen, contrast and clear the images so that the edge detection and measurement algorithm would better recognize the individual photos and extract them for comparison.
After the extraction, a face recognition software was implemented, based on artificial neural network models that are capable of self-optimizing their own pattern recognition processes while going through more and more data.