muellmat / OctoEye

Pupil tracking algorithm especially for the Octopus® 900 perimeter (Haag-Streit).

OctoEye

Pupil tracking algorithm especially for the Octopus® 900 perimeter (Haag-Streit).

Dependencies

• JDK 1.7
• OpenCV 2.4.9

How to run the example

1. Extract example files: cd data && tar xf 1.tar.xz
2. Download OpenCV. Create a new folder called lib.
   1. Windows: copy opencv-249.jar (including both x64 and x86 subdirectories) from path-to-opencv/build/java into your lib folder.
   2. Mac OS X: todo
   3. Linux: todo
3. Compile the source files. I'd suggest to create a new project using an IDE such as IntelliJ
   1. In the new project, open File > Project structure and add opencv-249.jar to the libraries.
   2. Edit run configuration. You will need a VM option: e.g. -Djava.library.path="lib/x86". Add program arguments: data/1.
4. Run the Main class.

How does the algorithm work?

todo

FAQ

1. What are these raw files?

   The Octopus 900 perimeter returns grayscale images as a byte sequence with a dimension of 320x240 px. So the raw files are just byte[] foo = new byte[320*240*1] saved to a file.

2. What is t, d, a and b in the screenshot?

   • t: needed time
   • d: pupil diameter (okay, actually it's 2*pupilMajorAxis)
   • a: major axis of the pupil ellipse (red)
   • b: minor axis of the pupil ellipse (blue)