ContextCam

Featured on Indico's Inspiring Hacks: Blog Post

ContextCam is a project from HackPrinceton Spring 2017. ContextCam runs on the Raspberry PI with a RPi Cameram, a single push button, and a mobile power source. It takes a picture, runs it through Watson and Indico to get context about the image. Then, through use of nltk and PyDictionary, it generates a description about the image and outputs as audio.

The positioning of the object is important, preferably in the center with a monochromatic background for better results. It currently takes top three predictions and forms a description with them. For emotional analysis, it gets the top three emotions and describes the person.

Demo

Demo

Project at devpost

Code Structure

• main.py The main process, does API calls, receive button and camera inputs, process labels, grammar, outputs to audio. All dependencies can be found at the top of the file to download.
• settings.py Not present but has the Watson BlueMix API authentication.
• .indicorc Also not present, stored in home directory and has Indico API authentication.

Reference

1. Bird, Steven, Edward Loper and Ewan Klein (2009), Natural Language Processing with Python. O’Reilly Media Inc.