Adaptive Manifolds for Real-Time High-Dimensional Filtering

Application Examples

Eduardo S. L. Gastal
and Manuel M. Oliveira

For up-to-date information refer to our Project Website:

This document contains various examples illustrating some of the possible applications of our adaptive-manifold high-dimensional filter. Click on any image to open it in its original size.

Important: These materials are best viewed in a web browser that is not Internet Explorer (IE), since it cannot resize the images properly. Furthermore, in IE you might need to allow ActiveX content to display the images.

Click on any image to open it in its original size.

Edge-Aware Smoothing

Non-Local Means Denoising

Denoising with Additional Information

Detail Enhancement

Indirect Illumination Filtering

For these examples filtering was performed in a 8-D dimensional space, composed of 3-D scene position, 3-D normal vectors, and 2-D image space coordinates. The standard deviations of the filter were SIGMA_S = 32, SIGMA_NORMALS = 0.2, SIGMA_POSITIONS = 0.2. The scene lies inside a [-1,1]^3 cube.