Abstract
Instagram is a popular image-sharing platform available for mobile devices that allows users to capture pictures, apply one of several artistic filters, and post the resultant images online for others to view. These filters are presented as black-box systems, providing no user-definable parameters or configurations, save the input image itself. In this exercise, we recreated the Instagram filter “Hefe.”
To achieve this, we generated a set of test images, applied the filter to each one in the Instagram app, and compared the pre- and post-filter images quantitatively. As a result of this exercise, we distilled our work down to a simple function that replicates the “Hefe” filter when applied to an image array. We were also able to create an inverse filter which, given an Instagram-processed image, was able to recover the original data. The images produced by these functions were accurate to a high degree of similarity.
Download Full Paper