Generalized multi-scale stacked sequential learning for multi-class classification

Puertas, Eloi; Escalera, Sérgio; Pujol, Oriol

doi:10.1007/s10044-013-0333-y

Cited by 14 publications

(9 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…One way to address this difficulty is modifying the neighborhood function and replacing the base classifiers used in the original MS‐SSL scheme by others capable of dealing with data belonging to N classes. This adaptation, called MMSSL , is presented and applied for the resolution of several multi‐class sequential learning problems in . In this study, the authors concluded that MMSSL shows significant performance improvement compared with classical approaches.…”

Section: Stacking Variants and Related Approachesmentioning

confidence: 88%

Generating ensembles of heterogeneous classifiers using Stacked Generalization

Sesmero

Ledezma

Sanchís

2015

WIREs Data Min & Knowl

128

View full text Add to dashboard Cite

Over the last two decades, the machine learning and related communities have conducted numerous studies to improve the performance of a single classifier by combining several classifiers generated from one or more learning algorithms. Bagging and Boosting are the most representative examples of algorithms for generating homogeneous ensembles of classifiers. However, Stacking has become a commonly used technique for generating ensembles of heterogeneous classifiers since Wolpert presented his study entitled Stacked Generalization in 1992. Studies that have addressed the Stacking issue demonstrated that when selecting base learning algorithms for generating classifiers that are members of the ensemble, their learning parameters and the learning algorithm for generating the meta-classifier were critical issues. Most studies on this topic manually select the appropriate combination of base learning algorithms and their learning parameters. However, some other methods use automatic methods to determine good Stacking configurations instead of starting from these strong initial assumptions. In this paper, we describe Stacking and its variants and present several examples of application domains.

show abstract

Section: Stacking Variants and Related Approachesmentioning

confidence: 88%

Generating ensembles of heterogeneous classifiers using Stacked Generalization

Sesmero

Ledezma

Sanchís

2015

WIREs Data Min & Knowl

128

View full text Add to dashboard Cite

show abstract

“…However, instead of just relying on individual descriptions, we exploit the confidences provided by the GMMs in the different cells and types of description altogether. This approach follows the Stacked Learning scheme (Cohen, 2005;Puertas et al, 2013), which involves training a new learning algorithm by combining previous predictions obtained with other learning algorithms. More precisely, each grid r is represented by a vector v r of confidences:…”

Section: Learning-based Fusion Approachmentioning

confidence: 99%

Multi-modal RGB–Depth–Thermal Human Body Segmentation

et al. 2016

Self Cite

View full text Add to dashboard Cite

This work addresses the problem of human body segmentation from multi-modal visual cues as a first stage of automatic human behavior analysis. We propose a novel RGB-Depth-Thermal dataset along with a multi-modal segmentation baseline. The several modalities are registered using a calibration device and a registration algorithm. Our baseline extracts regions of interest using background subtraction, defines a partitioning of the foreground regions into cells, computes a set of image features on those cells using different state-of-the-art feature extractions, and models the distribution of the descriptors per cell using probabilistic models. A supervised learning algorithm then fuses the output likelihoods over cells in a stacked feature vector representation. The baseline, using Gaussian Mixture Models for the probabilistic modeling and Random Forest for the stacked learning, is superior to other state-of-the-art methods, obtaining an overlap above 75% on the novel dataset when compared to the manually annotated ground-truth of human segmentations.

show abstract

“…where C c refers to the confidence map for the class c. The vector a is the resulting feature vector for p. In effect, this decomposition creates a feature vector that is linear in size with the number of classes, e.g kn, as a consequence the framework is suited for problems with a relatively low number of classes, although it can be compressed when dealing with a large number of classes (Puertas et al, 2015). Notice that the points in P i are found as the points with an Euclidean distance from the query point p falling in the interval i.…”

Section: Multi-scale Sequential Stacked Classifiermentioning

confidence: 99%

“…In (Puertas et al, 2015) they present the Multi-class Multiscale Stacked Sequential Learning (MMSSL) framework. The idea is to stack subsequent classifiers and cumulatively extend the feature sets with filtered versions of each classifiers predictions.…”

Section: Introductionmentioning

confidence: 99%