SMURFS: superpixels from multi-scale refinement of super-regions

Recent applications in computer vision have come to rely on superpixel segmentation as a pre-processing step for higher level vision tasks, such as object recognition, scene labelling or image segmentation. Here, we present a new algorithm, Superpixels from MUlti-scale ReFinement of Super-regions (S...

Full description

Bibliographic Details
Main Authors: Luengo, Imanol, Basham, Mark, French, Andrew P.
Format: Conference or Workshop Item
Published: 2016
Subjects:
Online Access:https://eprints.nottingham.ac.uk/37025/
_version_ 1848795377245880320
author Luengo, Imanol
Basham, Mark
French, Andrew P.
author_facet Luengo, Imanol
Basham, Mark
French, Andrew P.
author_sort Luengo, Imanol
building Nottingham Research Data Repository
collection Online Access
description Recent applications in computer vision have come to rely on superpixel segmentation as a pre-processing step for higher level vision tasks, such as object recognition, scene labelling or image segmentation. Here, we present a new algorithm, Superpixels from MUlti-scale ReFinement of Super-regions (SMURFS), which not only obtains state-of-the-art superpixels, but can also be applied hierarchically to form what we call n-th order super-regions. In essence, starting from a uniformly distributed set of super-regions, the algorithm iteratively alternates graph-based split and merge optimization schemes which yield superpixels that better represent the image. The split step is performed over the pixel grid to separate large super-regions into different smaller superpixels. The merging process, conversely, is performed over the superpixel graph to create 2nd-order super-regions (super-segments). Iterative refinement over two scale of regions allows the algorithm to achieve better over-segmentation results than current state-of-the-art methods, as experimental results show on the public Berkeley Segmentation Dataset (BSD500).
first_indexed 2025-11-14T19:31:07Z
format Conference or Workshop Item
id nottingham-37025
institution University of Nottingham Malaysia Campus
institution_category Local University
last_indexed 2025-11-14T19:31:07Z
publishDate 2016
recordtype eprints
repository_type Digital Repository
spelling nottingham-370252020-05-04T18:11:33Z https://eprints.nottingham.ac.uk/37025/ SMURFS: superpixels from multi-scale refinement of super-regions Luengo, Imanol Basham, Mark French, Andrew P. Recent applications in computer vision have come to rely on superpixel segmentation as a pre-processing step for higher level vision tasks, such as object recognition, scene labelling or image segmentation. Here, we present a new algorithm, Superpixels from MUlti-scale ReFinement of Super-regions (SMURFS), which not only obtains state-of-the-art superpixels, but can also be applied hierarchically to form what we call n-th order super-regions. In essence, starting from a uniformly distributed set of super-regions, the algorithm iteratively alternates graph-based split and merge optimization schemes which yield superpixels that better represent the image. The split step is performed over the pixel grid to separate large super-regions into different smaller superpixels. The merging process, conversely, is performed over the superpixel graph to create 2nd-order super-regions (super-segments). Iterative refinement over two scale of regions allows the algorithm to achieve better over-segmentation results than current state-of-the-art methods, as experimental results show on the public Berkeley Segmentation Dataset (BSD500). 2016-09-20 Conference or Workshop Item PeerReviewed Luengo, Imanol, Basham, Mark and French, Andrew P. (2016) SMURFS: superpixels from multi-scale refinement of super-regions. In: British Machine Vision Conference (BMVC 2016), 20-22nd Sept 2016, York, UK. Segmentation Super pixels http://www.bmva.org/bmvc/2016/papers/paper004/index.html
spellingShingle Segmentation
Super pixels
Luengo, Imanol
Basham, Mark
French, Andrew P.
SMURFS: superpixels from multi-scale refinement of super-regions
title SMURFS: superpixels from multi-scale refinement of super-regions
title_full SMURFS: superpixels from multi-scale refinement of super-regions
title_fullStr SMURFS: superpixels from multi-scale refinement of super-regions
title_full_unstemmed SMURFS: superpixels from multi-scale refinement of super-regions
title_short SMURFS: superpixels from multi-scale refinement of super-regions
title_sort smurfs: superpixels from multi-scale refinement of super-regions
topic Segmentation
Super pixels
url https://eprints.nottingham.ac.uk/37025/
https://eprints.nottingham.ac.uk/37025/