This project is the implementation of the UI part of my master thesis, Gotcha: Document Retireval System Using Visual Layout Manipulation. The purpose of my research is to allow users to search documents without the the need of typing keywords. The users can input a visual query based on document's visual layout, and the system will find the most closely matched document pages based on the difference of spatial relations and geometrical differences.
https://www.youtube.com/watch?v=jikMlTSqFIo&list=UUCMFdAMeq1t8v7MCakzR08g
- Ability to code in Ojective-C and create Mac OSX app
- Computer Vision (For document segmentation, crop out the parts with content)
- Machine Learning (For classifying what's inside the content blocks into text, title, image, etc.)