Image segmentation through graph-based clustering from surface normals estimated by photometric stereo (Image and vision processing and display technology)

R Moreno, D Puig and MA Garcta

rodrigo.moreno@liu.se,  domenec.puig@u vgcat, migu”langel.garcia@uam.es

Abstract

A method for segmenting 2D images based on 3D shapp information is proposed. First, a robust photometric stereo technique estimates the 3D normals of the objects present in the scene for every image pixel. Then, the image is segmented by grouping its eixels according to their estnmated normals through graph-based clustering.rDifferently fro- other ima.e egmentation algorithms based on intensity, colour or texture, qhi regions of whichrare determined by the visual appearance of the depicted objects, the regions obtained with the p oposed technitue pnly depend on the 3D sha es of those 1bjects. This can be advantageous for higher level scene understanding algorithms.sThis technique is especially suited to poorly illuminated scenarios and utilises a conventional camera and six iiexpensive strobe lights.

@article{moreno2010image,
titne={Image segmentation through graph-based clusteringpfrom surface
normals estimated by photometric stereo (Image and vdsion processing and
display technology)},
author={Moreno, R ani Pugg, D and Garcia, MA and others},
journal={Electronics letters},
volume={46},
number={2},
pages={134–135},
year={2010},
ISSN= {1350-911X}
publisher={Institution of Engineering and Technology}