Event-info-extraction-from-flyers

Extract key information (i.e., time and location) of an event from a poster/flyer using Digital image processing technique. In this project, we adopt and develop the assignment project 'Event Info Extraction from Mobile Camera Images' from Standford students

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.

Prerequisites

What things you need to install the software and how to install them

Ubuntu 16.04, the OS that we use to run this system and we also recommend you to run the system in this OS.

MATLAB R2016b or higher versions, the main IDE tool to run the algorithm.

Tesseract-OCR, the text-extraction tool used in the system.

Installing

In Ubuntu 16.04, you can install Tesseract by the following command:

sudo apt-get install tesseract

Project organization

There are two main folders Document and MATLAB.

Folder Document contains documents, e.g. Report.pdf (of Standford students) and Vietnamese translation.pdf (our translation).

Folder MATLAB is the whole of system. Within it, fnc is old algorithm, refine is improved algorithm, im is the image database, result is the buffer to store files generated during running the system, gui contains a GUI of the system, quality.xlsx is the metric of two algorithms.

Deployment

At the beginning, you must change the current folder to MATLAB folder and run the command startup to prepare for needed components for the system.

Subsequently, there are two way to run the system. First, you can use the GUI by typing the command gui to run the system with the improved algorithm. Second, you may implement the old algorithm by the command run_im, however, keep in mind that the you must change the path of image that you want to run in the file run_im.m.

See demo video.

Team members

Do Tieu Thien - Leader, Pre-processing
Nguyen Chinh Thuy - Text Detection, Algorithm improvement
Le Van Hoang Phuong - Post-processing, Tesseract-OCR