In this work, we propose an original segmentation model based on a
preliminary spatially adaptive non-linear dimensionality data
reduction step integrating contour and texture cues. This new
dimensionality reduction model aims at converting an input texture
image into a noisy color image in order to greatly simplify its
segmentation. In this de-texturing model, the spatially adaptive
non-local constraints based on edge and contour cues allows us to
efficiently regularize the resulting de-textured color image and
to efficiently combine region and edge cues in a data fusion model
used as data pre-processing step in a segmentation model.