Fast face and saliency aware collage creation for mobile phones

Abstract

This demonstration is of an automated method for creating image collage in a mobile phone. The app creates semantically meaningful collages from images based on faces, image saliency and hybrid blending. The algorithm is designed for computational efficiency inorder for it run on a mobile device. Due to the increase in the use of social networks, users are uploading images and videos from social events. Collage presents a useful crisper summary of the event. It is popular among users as evident from Layout from Instagram: Collage app, which has over 100 million downloads. A limitation of these apps is that they are dependent on the user for selecting the grid layout and optimal cropping of the images. Our smart collage over comes these limitations and to the best of our knowledge this is one of the first mobile based solutions, which considers the faces and the region importance and later merges the salient images automatically in a non-rigid grid layout. Other interesting works either have a semi-rigid [1] or a fully rigid grid based layout [2] or are desktop based [3].