written by Niru Maheswaranathan ([email protected])

This toolbox contains some useful scripts for generating and processing image ensembles in Matlab.

1. This toolbox uses images from the van Hateren natural image database. Download the images from the Bethge lab. A mirror is also hosted by Paul Ivanov at his website.
2. Specify the path to the image set you want to use (either the linearized .iml images or the deblurred, or calibrated, .imc images) in setupIPT.m
3. Make sure to run setupIPT.m before using the toolbox. You can add the entire toolbox to your path using pathtool in Matlab.

To use the toolbox, you'll need a copy of Matlab. I haven't figured out exactly what versions are supported yet. It might also work with Octave.

Secondary helper functions included in the helper_functions folder are not included in this list.

• setupIPT.m Sets up some paths for the imageprocessing toolbox (IPT). Run this before using the toolbox!
• generate_samples.m Generates (or samples) image patches from different model classes (gaussian noise, pink noise, natural images, or the dead leaves model)
• loadimage.m Loads an image from the van Hateren database
• oneoverf.m Generates 1/f noise (from Lawrence Cormack)
• plot_patches.m Plots rows of a matrix as images in a grid. Useful for visualizing the output of generate_samples.m
• downsample2.m 2D downsampling function
• upsample2.m 2D upsampling function
• gabor.m Generates a 2D gabor function

This code is provided for non-commercial research use only. There are no guarantees that anything works, but let me know if something doesn't work and I will fix it.