You have chosen to sponsor your bid up to a maximum amount of .
We are looking for a Java developer to create an application able to group together image and video files based upon similarities. The module will accept a source directory containing various multimedia files and compare the visual and audial features in each, compared to all other files in the directory. The application will also have a second ability to search for files similar to a selected inputted file.
Please note: This position is similar to the other advertised position, but does not include audio sorting.
To application will need to be able to sort and search files based upon the following techniques:
1. Joint-Histogram analysis (http://www.cs.cornell.edu/~rdz/papers/pz-jms99.pdf)
2. Fuzzy hashing
A Java application able to:
• Sort a directory (and files in enclosed subdirectories) of multimedia files based on similarity to each other file in the folder. The results will show batch groups of similar files together one after another.
• Search for similar files to a single inputted file
• Output the results of the sort using one of the following 3 methods (selected by the user):
1. By renaming the multimedia files with a select prefix (e.g. changing test.avi to 01-test.avi).
2. By creating an XML file showing the file details and the order/position in which it appears once sorted in select tags (e.g. <new order>1</new order>).
3. By adjusting a pre-exiting XML file which current notes the order of the file. In this instance a copy of the original XML will be made and the tag that contains the old file order will need to be updated. Sample XML will be sent to shortlisted candidates.
• Complete the sort/search and output the results as fast as possible (must include multithreading)
• Work as accurately as possible
• Work efficiently
• Be run via a very simple GUI (design of which will be sent to shortlisted candidates), or through an API
- Must work on Windows, Mac and Linux
- Be developed in Java and able to be integrated into further scripts and applications
- Must perform efficiently, and with a good accuracy (including multithreading)
- Must be able to handle the sort and search of millions of files in the same scan
- All parts of this development must be free for commercial usage
- A simple usage guide, along with a selection of test data must be supplied to the project owner
- Must fast scan video (i.e. so it does not need to be played in real-time)
- Able to sort all common multimedia file formats
- All parts of this project must comply with applicable UK law
Please note fuller details will be sent to shortlisted candidates (GUI design, XML layout, sample files etc). Please feel free to suggest alternative sorting techniques, but these must be discussed and agreed by the project owner.