RISF: Recursive Iterative Spanning Forest for Superpixel Segmentation

Galvão, Felipe Lemes; Falcão, Alexandre X.; Chowdhury, Ananda S.

doi:10.1109/sibgrapi.2018.00059

Cited by 8 publications

(16 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For that, we propose a hierarchical version of the Iterative Spanning Forest (ISF) superpixel segmentation framework [15]. This part is an extension of our previous work [18]. Here we show that our previous experimental settings were insufficient for a practical hierarchy 45 with more than two observation scales 3 , and present a more suitable alternative.…”

Section: Introductionmentioning

confidence: 92%

“…When G is a RAG, any pixel-based sampling strategy (including regular grid sampling) can be trivially used by converting each selected pixel to the region (node in the RAG) the pixel is part of, with the disadvantage that duplicates would have to be removed. We still use this adaptation of regular grid sampling in our proposed method as it was empirically observed that the amount of duplicates is small; an alternative that samples directly over the RAG was presented in [18].…”

Section: Iterative Spanning Forest (Isf)mentioning

confidence: 99%

“…Lastly, the convergence criteria (Line 5) is simplified to a fixed number of T = 10 iterations. For an in-depth description including the IFT algorithm and its sub-components, see [18]. The definition of a RISF method consists of defining multiple ISF methods with the added restriction that the target number of regions must be decreasing.…”

Section: Iterative Spanning Forest (Isf)mentioning

confidence: 99%

“…We also consider the variant RISF-old based on the experimental settings used in [18]; its difference to RISF is that it uses the same path-cost function and seed recomputation procedure as ISF in all scales. Additionally, the factor of reduction after each scale is varied to evaluate the performance of RISF-old when building hierarchies with different number of observation scales.…”

Section: Experimental Settingsmentioning

confidence: 99%

“…Superpixel algorithms were originally designed to create an image partition from its pixel set [13,14,15]. Some of these algorithms can be easily executed recursively on each current image partition to create the next coarser image partition for hierarchical image segmentation [16,17,18]. This allows the addition of a new coarser partition to the hierarchy at each iteration (see Figure 1), generally with significantly less number of regions than the previous one and overall creating a relatively small number of observation scales 2 .…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Image segmentation using dense and sparse hierarchies of superpixels

Galvão

Guimarães

Falcão

2020

Pattern Recognition

View full text Add to dashboard Cite

We investigate the intersection between hierarchical and superpixel image segmentation. Two strategies are considered: (i) the classical region merging, that creates a dense hierarchy with a higher number of levels, and (ii) the recursive execution of some superpixel algorithm, which generates a sparse hierarchy with fewer levels. We show that, while dense methods can capture more intermediate or higher-level object information, sparse methods are considerably faster and usually with higher boundary adherence at finer levels. We first formalize the two strategies and present a sparse method, which is faster than its superpixel algorithm and with similar boundary adherence.We then propose a new dense method to be used as post-processing from the intermediate level, as obtained by our sparse method, upwards. This combination results in a unique strategy and the most effective hierarchical segmentation method among the compared state-of-the-art approaches, with efficiency comparable to the fastest superpixel algorithms.

show abstract

Section: Introductionmentioning

confidence: 92%

Section: Iterative Spanning Forest (Isf)mentioning

confidence: 99%

Section: Iterative Spanning Forest (Isf)mentioning

confidence: 99%

Section: Experimental Settingsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Image segmentation using dense and sparse hierarchies of superpixels

Galvão

Guimarães

Falcão

2020

Pattern Recognition

View full text Add to dashboard Cite

show abstract

Fast and Effective Superpixel Segmentation Using Accurate Saliency Estimation

Belém

Borlido

João

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Novel Arc-Cost Functions and Seed Relevance Estimations for Compact and Accurate Superpixels

et al. 2023

View full text Add to dashboard Cite

On the verge of superpixel methods exploiting saliency information, the Superpixels through Iterative CLEarcutting (SICLE) framework has reported fast and accurate superpixel delineation. It is composed of three steps: (i) seed oversampling; (ii) superpixel generation; and (iii) seed removal. It starts from (i) and applies several iterations of (ii) and (iii) until reaching the desired superpixel quantity. In this work, we improve SICLE such that it can now generate compact superpixels with accurate delineation. We exploit differential computation and propose several novel functions for steps (ii) and (iii) for proper saliency incorporation, compact superpixel generation, and improvement in speed and delineation. Results show that, with our proposals, SICLE achieves state-of-the-art performance in delineation and speed whenever saliency is absent with on-par compacity. When an accurate saliency map is provided, its performance improves significantly and requires only two iterations for segmentation.

show abstract

RISF: Recursive Iterative Spanning Forest for Superpixel Segmentation

Cited by 8 publications

References 21 publications

Image segmentation using dense and sparse hierarchies of superpixels

Image segmentation using dense and sparse hierarchies of superpixels

Fast and Effective Superpixel Segmentation Using Accurate Saliency Estimation

Novel Arc-Cost Functions and Seed Relevance Estimations for Compact and Accurate Superpixels

Contact Info

Product

Resources

About