|Title:||Automated image analyzing and categorizing program|
MostPhotos is a company that buys and sells images via our Internet website. Every day we receive thousands of new photographs and illustrations. In order for our customers to quickly find what they are looking for, search functionality is used with the different tags of images. Because customers should be able to search for photos – and illustrations specifically – and since the company is based on automation, a program that can tell the difference between such images with high accuracy is required.
Tobias Carlsson has earlier been developing such a program in collaboration with a project group in a previous course at MDH (DVA313 – Software Engineering). This basic version of the program exists as a Windows application written in C#. There are still many improvements to be done if this program is to meet the full requirements of our customers however.
The goal of the project is to improve the existing program engine to that of the desired functionality, and to port the system to C++ code.
Potential improvements may include:
* Significantly decrease the execution time
* General code optimization
* Add (at least one more) algorithm to increase accuracy
* Refinement of the existing algorithms
* Create a training engine for the algorithms
* The final analyze should be performed on the entire image area
* A larger sample volume should be used in order to calibrate the algorithms correctly
The program shall be implemented using GraphicsMagick or any other suitable graphical framework and shall be delivered as a separate module with a well defined input and output which could easily be integrated in our existing system. The final product shall also include instructive documentation and a test report.
|Company:||MostPhotos, kontaktperson: Andreas Andersen|
|Prel. end date:||2011-06-10|
|Student:||Tobias Carlsson firstname.lastname@example.org|