The PunMe app allows a user to take a photo or upload a photo and receive back a pun generated based on the subject of the photo.

The mobile application, built for iOS iPhone, iTouch and iPad in Swift 3, calls our API using the Alamofire networking framework and sends a photo in the form of a jpeg image, which the user can either take directly or select from his/her camera roll. From there, the app receives a JSON with a pun and a key word which appears in the pun, which is then displayed to the user along with the image.

Our backend server is a RESTful API built in Java Spring and deployed to Amazon Web Services using Boxfuse. When an image is received, it calls Microsoft's Computer Vision API which returns a caption describing the contents of the image, among other information about the image. We then use Google's Cloud Natural Language Processing API for syntax analysis of the caption. After this, we perform a variety of custom checks to determine key subjects of the image input by the user, and consequently the direct subject. We then use this word to search punoftheday.com for puns relating to the subject of the image.

PunMe is available at punme.net.