Semi-supervised hyperspectral pixel classification using interactive labeling

Rajadell, Olga; García-Sevilla, Pedro; Dinh, Viet Cuong; Duin, Robert P. W.

doi:10.1109/whispers.2011.6080905

Cited by 8 publications

(6 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…On the other hand, in segmentation and classification of this kind of images the training data used has not been a concerned so far, without worrying about providing the most reliable information (Comaniciu and Meer, 2002). The scheme suggested in (Rajadell et al, 2011) was a first attempt in this sense. It was proposed an unsupervised selection of the training samples based on the analysis of the feature space to provide a representative set of labeled data.…”

Section: Preliminariesmentioning

confidence: 99%

“…In (Rajadell et al, 2011) it was suggested to weigh the spatial coordinates by an arbitrary number to reinforce two samples that are close spatially to have a closer distance and the way round. Such a weight should be decided in terms of the range of the features provided by the spectrometer so the coordinates are overweighed but they do not cause the rest of features be dismissed in the global measure.…”

Section: Spatial Improvementmentioning

confidence: 99%

“…We suggest to use a clustering analysis to find samples of interest, ask an expert for their labels and classify using that labeled set obtained. This scheme was presented in (Rajadell et al, 2011) where a KNN1 classifier was used. Here we suggest to assign labels to unlabeled samples according to the result given by the cluster itself and the labels provided by the expert for the modes of clusters.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Automatic Selection of the Training Set for Semi-Supervised Land Classification and Segmentation of Satellite Images

Rajadell

García-Sevilla

2012

Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods

View full text Add to dashboard Cite

Different scenarios can be found in land classification and segmentation of satellite images. First, when prior knowledge is available, the training data is generally selected by randomly picking samples within classes. When no prior knowledge is available the system can pick samples at random among all unlabeled data, which is highly unreliable, and ask the expert to label them or it can rely on the expert collaboration to improve progressively the training data applying an active learning function. We suggest a scheme to tackle the lack of prior knowledge without actively involving the expert, whose collaboration may be expensive. The proposed scheme uses a clustering technique to analyze the feature space and find the most representative samples for being labeled. In this case the expert is just involved in labeling once a reliable training data set for being representative of the features space. Once the training set is labeled by the expert, different classifiers may be built to process the rest of samples. Three different approaches are presented in this paper: the result of the clustering process, a distance based classifier, and support vector machines (SVM).

show abstract

Section: Preliminariesmentioning

confidence: 99%

Section: Spatial Improvementmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Automatic Selection of the Training Set for Semi-Supervised Land Classification and Segmentation of Satellite Images

Rajadell

García-Sevilla

2012

Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods

View full text Add to dashboard Cite

show abstract

“…However, in segmentation and classification of this kind of images, training sets are often built by randomly picking a percentage, against the principle of providing more reliable information. Here the proposed semi-supervised pixel classification scheme presented Rajadell et al 4 is explained. The scheme makes an unsupervised selection of the training samples based on the analysis of the feature space.…”

Section: Classification Schemementioning

confidence: 99%

“…Consequently spatial criteria added in two different ways will help to make a better selection and provide a much lower error than a random selection and still outperform the ones presented in previous work. 4 This spatial criterion can be used after the clustering algorithm or within. In the first case, we discard selected samples within a given neighbourhood to get a smaller set of samples that spatially represent the same region in the image.…”

Section: Introductionmentioning

confidence: 99%

Selection of samples for active labeling in semi-supervised hyperspectral pixel classification

et al. 2011

Self Cite

View full text Add to dashboard Cite

One of the problems in semi-supervised land classification tasks lies in improving classification results without increasing the number of pixels to be labeled. This would be possible if, instead of increasing the amount of data we increased the reliability of the data. We suggest to replace the random selection by a unsupervised clustering based selection strategy in building the training data. We use a mode seeking clustering method to search for cluster representatives, which will be labeled and then used for training. Here an improvement to the result of the clustering algorithm is introduced by taking advantage of the spatial information in the image. The number of selected samples provided by the clustering can be reduced by using a spatial-density criterion to dismiss redundant training information. Two different alternatives are considered for a spatial criterion, one dismisses selected samples in the same neighbourhood and the other includes the pixel coordinates for giving the spatial information a larger weight in the clustering. Both alternatives improve the classification-segmentation results. The classification scheme with training selection provides state-of-the-art pixel classification results using a smaller training set and suggests an alternative to random selection.

show abstract

Exploiting Spatial Correlation of Spectral Signature for Training Data Selection in Hyperspectral Image Classification

Appice

Guccione

2016

Discovery Science

View full text Add to dashboard Cite

Semi-supervised hyperspectral pixel classification using interactive labeling

Cited by 8 publications

References 11 publications

Automatic Selection of the Training Set for Semi-Supervised Land Classification and Segmentation of Satellite Images

Automatic Selection of the Training Set for Semi-Supervised Land Classification and Segmentation of Satellite Images

Selection of samples for active labeling in semi-supervised hyperspectral pixel classification

Exploiting Spatial Correlation of Spectral Signature for Training Data Selection in Hyperspectral Image Classification

Contact Info

Product

Resources

About