Efficient High-Dimensional, Edge-Aware Filtering
Eduardo S. L. Gastal · IEEE Computer Graphics and Applications · 2016
Traditional computer graphics filters are agnostic to the concept of objects and the transition between objects (edges), but image filters need to take edges into account in order to produce high-quality, artifact-free, beautifully formed images. Eduardo S.L. Gastal's dissertation focuses on efficient algorithms for performing edge-aware image and video filtering. The techniques he describes here leverage the known relationship between edge-aware filtering and high-dimensional spaces and are capable of achieving real-time performance in high-resolution data.