Image Processing with Natural Scene Statistics

This site is a free service provided by the Center for Perceptual Systems at the University of Texas at Austin.

At the Center for Perceptual Systems, we measure the statistical properties of images by analyzing very large sets of natural images. Among other applications, these statistical measurements can be used to perform digital image processing tasks such as enlargement (super-resolution), denoising, deblurring, color filter array interpolation, and image compression. This site demonstrates some of these applications by allowing users to upload images and get results in real time. These applications use a computational method called recursive condition means (RCM). The site also provides a platform for researchers, photographers, and algorithm developers to compare methods.

Current site applications include

• Denoising
• Enlargement
• Realistic camera noise synthesis

More technical information may be found at the Natural Scene Statistics in Vision Science website.