VizPol: Real-Time Symbol Recognition for Field Reporting

Screen Shot 2020-12-03 at 3.01.23 PM.png

Video


Team Information

Team Members

  • Ishaan Jhaveri, Computational Research Fellow, Tow Center for Digital Journalism, Graduate School of Journalism , Columbia University

  • Nina Berman, Professor of Journalism, Tow Center for Digital Journalism, Graduate School of Journalism , Columbia University

  • Susan McGregor, Associate Research Scholar and Co-Chair of the Data, Media & Society Center, The Data Science Institute, Columbia University

  • Shih-Fu Chang, Richard Dicker Professor, Digital Video and Multimedia (DVMM) Lab, Electrical Engineering Department, Computer Science Department, Columbia Engineering

Abstract

The VizPol app is a graphical-recognition system designed to support and inform the work of journalists as they encounter unfamiliar and evolving graphical imagery in the field. Using a combination of computer vision and machine learning techniques and user feedback, the mobile application allows users to select symbols within photographs on their mobile device and receive information about the meaning of those symbols in real-time. A web-based version of the tool allows newsroom editors and others the use the system on desktop devices. The goal is to provide journalists with more information about rapidly-evolving political symbols in a way that can be seamlessly integrated into their existing workflows.

Contact this Team

Team Contact: Ishaan Jhaveri (use form to send email)

Previous
Previous

VisiFit: Structuring Iterative Improvement for Novice Designers

Next
Next

Monte Carlo Tree Search for Generating Interactive Data Analysis Interfaces