I’ve finished my dissertation, this is it!

I chose to research into the area of automatic photograph orientation, with the aim of implementing an application to correctly orientate images of incorrect orientations. I achieved a mark of 78% and was awarded the first prize for a multimedia project from the year.

Automatic photograph orientation is the method of detecting the correct orientation of an image, through several possible feature extraction techniques, and rotating that image to be correctly orientated. I researched into this area, implementing several methods of feature extraction, to produce an application to correctly orientate images.

I produced a full report of my research and findings as well as details of the implementation of the application.


Automatic photograph orientation

The application I produced was developed using C++ and Cocoa with an Objective-C bridge between the back and front end. It implemented several feature extraction techniques and achieved a fairly high success rate with outdoor images; mainly due to a sophisticated colour feature detection algorithm.

The application interface
The application interface

I learnt how to keep focus and work on a long-term, large scale project. I also had to work independently across various disciplines. It has definitely been the biggest challenge I have faced at uni.

  • Delicious
  • Digg
  • Facebook
  • FriendFeed
  • LinkedIn
  • MySpace
  • Netvibes
  • NewsVine
  • Posterous
  • Reddit
  • StumbleUpon
  • Technorati
  • Tumblr
  • Twitter
  • Yahoo! Buzz
  • email
  • Print

No responses to “Automatic photograph orientation”

  1. No comments

Leave a reply