Two lattice computing approaches for the unsupervised segmentation of hyperspectral images

Graña, Manuel; Villaverde, Iván; Maldonado, José O.; Hernández, Carmen Añez

doi:10.1016/j.neucom.2008.06.026

Cited by 72 publications

(37 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…These ideas have been applied in diverse areas, such as pattern recognition (Ritter et al, 1998), associative memories in image processing (Ritter et al, 2003;Ritter & Gader, 2006;), computational intelligence (Graña, 2008), industrial applications modeling and knowledge representation (Kaburlasos & Ritter, 2007), and hyperspectral image segmentation (Graña et al, 2009;Ritter et al, 2009;Ritter & Urcid, 2010;.…”

Section: Lattice Algebra Operationsmentioning

confidence: 99%

Convex Set Approaches for Material Quantification in Hyperspectral Imagery

Valdiviezo-N.¹,

Urcid²

2012

Earth Observation

View full text Add to dashboard Cite

Section: Lattice Algebra Operationsmentioning

confidence: 99%

Convex Set Approaches for Material Quantification in Hyperspectral Imagery

Valdiviezo-N.¹,

Urcid²

2012

Earth Observation

View full text Add to dashboard Cite

“…The second algorithm tested is Endmember Induction Algorithm (EIHA) was fully described in [4], so that here we will only recall some of its main features. The algorithm is based on the equivalence between Strong Lattice Independence and Affine Independence [8].…”

Section: Endmember Induction Algorithmsmentioning

confidence: 99%

“…Because of the convergence properties of the Lattice AutoAssociative Memories, lattice dependent vectors will be recall-invariant, so lattice independent vectors can be detected as non-recall-invariant vectors. The EIHA proposed in [4] includes a noise filter that discards candidate vectors which are too close to the already detected endmembers. Figure 4 shows the precision k (H) and recall k (H) results of the N-FINDER and EIHA (denoted LAM in the figures) algorithms respect to three defined synthetic hyperspectral image databases, generated from a collection of 10 basic endmembers selected from the USGS library of spectral signatures, for all possible values of the size of the response K using the dissimilarity function 2.…”

Section: Endmember Induction Algorithmsmentioning

confidence: 99%

“…They try either to select some image pixel spectra as the best approximation to the endmembers in the image (i.e. [4]), or to compute estimations of the endmembers on the basis of the transformations of the image data (i.e. [5,10]).…”

Section: Introductionmentioning

confidence: 99%

“…We do not impose classification like schemes as the performance measures, but we evaluate the ability of the algorithms to uncover the underlying mixtures. We apply this methodology to compare the Endmember induction Heuristic Algorithm (EIHA) [4], with respect to the well known geometrical algorithm N-FINDER [10].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

On the Use of a Hybrid Approach to Contrast Endmember Induction Algorithms

Veganzones

Hernández

2010

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. In remote sensing hyperspectral image processing, identifying the constituent spectra (endmembers) of the materials in the image is a key procedure for further analysis. The contrast between Endmember Inductions Algorithms (EIAs) is a delicate issue, because there is a shortage of validation images with accurate ground truth information, and the induced endmembers may not correspond to any know material, because of illumination and atmospheric effects. In this paper we propose a hybrid validation method, composed on a simulation module which generates the validation images from stochastic models and evaluates the EIA through Content Based Image Retrieval (CBIR) on the database of simulated hyperspectral images. We demonstrate the approach with two EIA selected from the literature.

show abstract

Morphological Associative Memories

Valle

Sussner

Santos

2018

Wiley Encyclopedia of Electrical and Electronics Engineering

View full text Add to dashboard Cite

Morphological associative memories (MAMs) belong to the class of morphological neural networks that – in contrast to classical models of artificial neural networks – employ operators of mathematical morphology in order to aggregate the synaptic weights and the incoming signal. The original and most widely known MAM model as well as some recent modifications of this model can be formulated using matrix operation of a lattice algebra called minimax algebra that is closely related to the umbra approach of grayscale mathematical morphology. Fuzzy MAM models execute basic operations of fuzzy mathematical morphology in the recall phase. In this article, we briefly review the aforementioned MAM models and discuss some applications in classification, hyperspectral image analysis, and prediction.

show abstract

Two lattice computing approaches for the unsupervised segmentation of hyperspectral images

Cited by 72 publications

References 15 publications

Convex Set Approaches for Material Quantification in Hyperspectral Imagery

Convex Set Approaches for Material Quantification in Hyperspectral Imagery

On the Use of a Hybrid Approach to Contrast Endmember Induction Algorithms

Morphological Associative Memories

Contact Info

Product

Resources

About