Superpixels and Supervoxels in an Energy Optimization Framework

By Olga Veksler, Yuri Boykov, Paria Mehrani. European Conference on Computer Vision, 2010

512 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/ddd36382c3034cd9eebe18a61e8bedd2d1a40601

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

Geodesic Saliency Using Background Priors (2012, methodology): “For applications that require high accuracy such as object segmentation [17, 18], we propose to use superpixels [28] as image patches and an exact shortest path algorithm.”
Superpixels and Polygons Using Simple Non-iterative Clustering (2017, background): “[32], generate superpixels by stitching together overlapping image patches.”
Lazy Random Walks for Superpixel Segmentation (2014, background): “The second category of superpixel algorithms considers the compactness constrains, such as normalized cuts [7], lattice cut [16], TurboPixels [17], and graph cut [19] approaches.”
A systematic comparison between visual cues for boundary detection. (2016, unknown): “…color, texture, or motion attributes) to sequentially partition an image into regions (e.g., Felzenszwalb & Huttenlocher, 2004; Cremers, Rousson, & Deriche, 2006; Moore, Prince, Warrell, Mohammed, & Jones, 2008; Levinshtein et al., 2009; Veksler, Boykov, & Mehrani, 2010; Achanta et al., 2012).”
Piecewise Rigid Scene Flow (2013, methodology): “(8) alone, which amounts to the superpixelization approach of [21].”
Entropy rate superpixel segmentation (2011, methodology): “However, they produce superpixels with irregular sizes and shapes which tend to straddle multiple objects as pointed out in [10, 22].”
Superpixels: An evaluation of the state-of-the-art (2016, methodology): “Veksler et al. [36] distinguish superpixel algorithms from segmentation algorithms running in “oversegmenta-tion mode”.”
Voxel Cloud Connectivity Segmentation - Supervoxels for Point Clouds (2013, methodology): “To measure this quantitatively, we have used two standard metrics for boundary adherence- boundary recall and undersegmentation error[7, 13].”
SEEDS: Superpixels Extracted Via Energy-Driven Sampling (2012, methodology): “Veksler and Boykov [13] place overlapping patches over the image and assign each pixel to one of those by inferring a solution with graph-cuts.”
Beyond pixels: A comprehensive survey from bottom-up to semantic image segmentation and cosegmentation (2015, methodology): “Such type of segmentation methods with user inputs are termed as interactive image segmentation methods.”
ImageNet Auto-Annotation with Segmentation Propagation (2014, methodology): “A simple and widely used technique is to group pixels into superpixels [67,39], and assume that all the pixels inside a superpixel share the same label.”
Superpixel segmentation using Linear Spectral Clustering (2015, methodology): “According to Figure 3, EneOpt0 performs the worst in terms of boundary adherence among the five selected algorithms, probably because that it uses a variation of the minimum cut strategy which may suffer from the bias of cutting out small set of pixels, leading to under segmentation errors in practi”
3D Scene Flow Estimation with a Piecewise Rigid Scene Model (2015, methodology): “Note that a similar strategy was proposed by Veksler et al. (2010) to compute an over-segmentation of a single image.”
SLIC Superpixels Compared to State-of-the-Art Superpixel Methods (2012, methodology): “Ability to specify the amount of superpixels, control their compactness, and ability to generate supervoxels is also provided.”
Image Segmentation by Cascaded Region Agglomeration (2013, methodology): “SLIC and ERS have been shown in [1, 14] to outperform other methods, including MeanShift [4], Turbopixels [13], graph-based methods [18, 6, 22], and others [17, 8], on at least some of the data sets, and we do not compare to those methods.”