Optimum-Path Forest based on k-connectivity: Theory and applications

Papa, João Paulo; Fernandes, Silas Evandro Nachif; Falco, Alexandre Xavier

doi:10.1016/j.patrec.2016.07.026

Cited by 52 publications

(30 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…ISF falls in the category of graph-based algorithms as a particular case of a more general framework [11] -the Image Foresting Transform (IFT). The IFT is a framework for the design of image operators based on connectivity, such as distance and geodesic transforms, morphological reconstructions, multiscale skeletonization, image description, regionand boundary-based image segmentation methods [26], [24], [27], [28], [29], [30], [31], [32], [33], with extensions to clustering and classification [34], [35], [36], [37], [38]. As discussed in [39], by choice of the connectivity function, the IFT algorithm computes a watershed transform from a set of seeds that corresponds to a graph cut in which the minimum gradient value in the cut is maximized.…”

Section: Related Workmentioning

confidence: 99%

An Iterative Spanning Forest Framework for Superpixel Segmentation

Muñoz

Chowdhury

Alexandre

et al. 2019

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

Superpixel segmentation has become an important research problem in image processing. In this paper, we propose an Iterative Spanning Forest (ISF) framework, based on sequences of Image Foresting Transforms, where one can choose i) a seed sampling strategy, ii) a connectivity function, iii) an adjacency relation, and iv) a seed pixel recomputation procedure to generate improved sets of connected superpixels (supervoxels in 3D) per iteration. The superpixels in ISF structurally correspond to spanning trees rooted at those seeds. We present five ISF methods to illustrate different choices of its components. These methods are compared with approaches from the state-of-the-art in effectiveness and efficiency. The experiments involve 2D and 3D datasets with distinct characteristics, and a high level application, named sky image segmentation. The theoretical properties of ISF are demonstrated in the supplementary material and the results show that some of its methods are competitive with or superior to the best baselines in effectiveness and efficiency.

show abstract

Section: Related Workmentioning

confidence: 99%

An Iterative Spanning Forest Framework for Superpixel Segmentation

Muñoz

Chowdhury

Alexandre

et al. 2019

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

show abstract

“…A new classifier that has been highlighted in the literature is a technique called Optimum-Path Forest (OPF). The OPF proposes to classify patterns using graph theory concepts [18] [19]. This approach emerged as a generalization of the Image Foresting Transform (IFT) [17].…”

Section: Introductionmentioning

confidence: 99%

Remote Sensing Applied to the Extraction of Road Geometric Features Based on Optimum Path Forest Classifiers, Northeastern Brazil

Macedo¹,

Maia²,

Rabbani³

et al. 2020

JGIS

View full text Add to dashboard Cite

One of the principal difficulties related to road safety management in Brazil is the lack of data on road projects, especially those on rural roads, which makes it difficult to use road safety studies and models from other countries as a reference. Updating road networks through the use of hyperspectral remote sensing images can be a good alternative. However, accurately recognizing and extracting hyperspectral images from roads has been recognized as a challenging task in the processing of hyperspectral data. In order to solve the aforementioned challenges, Hyperion hyperspectral images were combined with the Optimum Forest Path (OPF) algorithm for supervised classification of rural roads and the effectiveness of the OPF and SVM classifiers when applied to these areas was compared. Both classifiers produced reasonable results, however, the OPF algorithm outperformed SVM. The higher classification accuracy obtained by the OPF was mainly attributed to the ability to better distinguish between regions of exposed soil and unpaved roads.

show abstract

“…Such samples try to conquer the remaining ones offering to them optimum-path costs, and when a sample is conquered, it receives the label of its conqueror. A new variant of the OPF classifier that makes use of a k-nearest neighborhood (k-nn) graph named OPF knn was proposed by Papa and Falcão [21,22,25], and its semi-supervised version has been presented by Amorim et al [4]. An interesting property stated by Souza et al [29] concerns that OPF is equivalent to 1-NN when all training samples are used as prototypes.…”

Section: Introductionmentioning

confidence: 99%

Improving optimum-path forest learning using bag-of-classifiers and confidence measures

Fernandes

Papa

2017

Pattern Anal Applic

Self Cite

View full text Add to dashboard Cite

Machine learning techniques have been actively pursued in the last years, mainly due to the great number of applications that make use of some sort of intelligent mechanism for decision-making processes. In this work, we presented an ensemble of optimum-path forest (OPF) classifiers, which consists into combining different instances that compute a score-based confidence level for each training sample in order to turn the classification process "smarter", i.e., more reliable. Such confidence level encodes the level of effectiveness of each training sample, and it can be used to avoid ties during the OPF competition process. Experimental results over fifteen benchmarking datasets have shown the effectiveness and efficiency of the proposed approach for classification problems, with more accurate results in more than 67% of the datasets considered in this work. Additionally, we also considered a bagging strategy for comparison purposes, and we showed the proposed approach can lead to considerably better results.

show abstract

Optimum-Path Forest based on k-connectivity: Theory and applications

Cited by 52 publications

References 22 publications

An Iterative Spanning Forest Framework for Superpixel Segmentation

An Iterative Spanning Forest Framework for Superpixel Segmentation

Remote Sensing Applied to the Extraction of Road Geometric Features Based on Optimum Path Forest Classifiers, Northeastern Brazil

Improving optimum-path forest learning using bag-of-classifiers and confidence measures

Contact Info

Product

Resources

About