Privacy mechanism for photo storage

Abstract

Photo-sharing has become an integral part of our daily lives. We willingly share photos online to pass on information, share about our lives and take part in online marketing activities. There is also a new trend of businesses engaging in digital marketing, by entailing its consumers to take photos of themselves with their products, in exchange for giveaway prizes. Furthermore, on Instagram alone, up to 95 million photos and videos are uploaded each day [24]. With this vast number of photos being shared daily on so many different platforms, one must consider the privacy aspect of such a phenomenon, especially since there is a tendency to neglect filtering our photos before uploading them on these platforms. This work aims to provide an alternative image privacy mechanism by first detecting object classes in an image, followed by applying a blurring filter onto privacy-sensitive object classes. Deep learning framework is applied for image segmentation on 21 object classes and image processing techniques are used for blurring the object classes.