#### DMCA

## Contents lists available at SciVerse ScienceDirect Expert Systems with Applications

### Citations

5951 | Neural Networks: A Comprehensive Foundation - Haykin - 1999 |

3749 | A modified particle swarm optimizer - Shi, Eberhart |

933 | A Direct Adaptive Method for Faster Backpropagation Learning: The RPROP Algorithm - Riedmiller, Braun - 1993 |

245 |
A review of target decomposition theorems in radar polarimetry.
- Cloude, Pottier
- 1996
(Show Context)
Citation Context ...a/A) decomposition employ the second order polarimetric representations of PolSAR data (such as covariance matrix or coherency matrix) to characterize distributed scatterers. The H/a/A decomposition (=-=Cloude & Pottier, 1996-=-) is based on eigen analysis of the polarimetric coherency matrix, h[T]i: hTi k1e1e T 1 þ k2e2e T 2 ð3Þ þ k3e3e T 3 ð5Þ where k1 > k2 > k3 P 0 are real eigenvalues and the corresponding orthonormal e... |

233 | A theory of networks for approximation and learning
- Poggio, Girosi
- 1992
(Show Context)
Citation Context ...recurrent ANN there can be backward loops in the network structure, while in feedforward ANNs such loops are not allowed. A popular type of feedforward ANN is the radial basis function (RBF) network (=-=Poggio & Girosi, 1989-=-), which has always two layers in addition to the passive input layer: a hidden layer of RBF units and a linear output layer. Only the output layer has connection weights and biases. The activation fu... |

209 |
An entropy based classification scheme for land applications of polarimetric SAR
- Cloude, Pottier
- 1997
(Show Context)
Citation Context ...ion. For multilook data represented in covariance or coherency matrices, Lee et al. (1999) proposed a new unsupervised classification method based on combination of polarimetric target decomposition (=-=Cloude & Pottier, 1997-=-) and the maximum likelihood classifier using the complex Wishart distribution. The unsupervised Wishart classifier has an iterative procedure based on the wellknown K-means algorithm, and has become ... |

102 |
Classification of multi-look polarimetric SAR imagery based on complex Wishart distribution
- Lee, Grunes, et al.
- 1994
(Show Context)
Citation Context ...d polarimetric SAR data, the three polarimetric components (HH, HV, and VV) has a multivariate complex Gaussian distribution and the complex covariance matrix form has a complex Wishart distribution (=-=Lee et al., 1994-=-). Due to presence of speckle noise and random vector scattering from surface or volume, polarimetric SAR data are often multilook processed by averaging n neighboring pixels. By using the Pauli based... |

93 |
An analysis of co-occurrence texture statistics as a function of grey-level quantization
- Clausi
- 2002
(Show Context)
Citation Context ...ttier decomposition includes entropy (H), anisotropy (A), estimates of the set of average angles (a, b, d, and c), three real eigenvalues (k1; k2; k3), and span. As suggested by the previous studies (=-=Clausi, 2002-=-; Zhang et al., 2009) appropriate texture measures for SAR imagery based on the gray level co-occurrence probabilities are included in the feature set to improve its discrimination power and classific... |

63 | Backpropagation: Theory and Architectures and Applications. Lawrence Erlbaum Associates: Boca - Chauvin, DE - 1995 |

56 | Polarimetric SAR speckle filtering and its implications for classification, - Lee, Grunes, et al. - 1999 |

54 |
Feature extraction from wavelet coefficients for pattern recognition tasks
- Pittner, Kamarthi
- 1999
(Show Context)
Citation Context ...pixel. Dimensionality reduction of input feature information improves efficiency of learning for a neural network classifier due to a smaller number of input nodes (to avoid curse of dimensionality) (=-=Pittner & Kamarthi, 1999-=-) and reduces computation time. For the purpose of normalizing and scaling the feature vector, each feature dimension is first normalized to have a zero mean and unity standard deviation before princi... |

45 |
Unsupervised classification of multifrequency and fully polarimetric SAR images based on the H/A Alpha - Wishart classifier,
- Ferro-Famil, Pottier, et al.
- 2001
(Show Context)
Citation Context ...s to color in this figure legend, the reader is referred to the web version of this article.) have been widely used in the polarimetric SAR literature over the last two decades (Ersahin et al., 2004; =-=Ferro-Famil, Pottier, & Lee, 2001-=-; Fukuda & Hirosawa, 1999), and distributed as multi-look processed and publicly available through the polarimetric SAR data processing and educational tool (PolSARpro) by ESA (The Polarimetric SAR Da... |

29 | Comparing Cooccurrence Probabilities and Markov Random Fields for Texture Analysis
- Clausi, Yue
- 2004
(Show Context)
Citation Context .../a/A decomposition based features combined with the backscattering power (Span), and the gray level co-occurrence matrix (GLCM) based texture features as suggested by the results of previous studies (=-=Clausi & Yue, 2004-=-; Ersahin, Scheuchl, & Cumming, 2004). The performance of the proposed RBF network based classifier is evaluated using the fully polarimetric San Francisco Bay and Flevoland data sets acquired by the ... |

24 |
Unsupervised classification scheme of PolSAR images based on the complex Wishart distribution and the entropy alpha anisotropy Polarimetric decomposition theorem. Proceeding of EUSAR
- Pottier, Lee
- 2000
(Show Context)
Citation Context ...ic SAR data. In the literature, the classification algorithms for polarimetric SAR can be divided into three main classes: (1) classification based on physical scattering mechanisms inherent in data (=-=Pottier & Lee, 2000-=-; van Zyl, 1989), (2) classification based on statistical characteristics of data (Lee et al., 1999; Wu, Ji, Yu, & Su, 2008) and (3) classification based on image processing techniques (Ince, 2010; Ta... |

20 |
A wavelet-based texture feature set applied to classification of multi-frequency polarimetric SAR images,
- Fukuda, Hirosawa
- 1999
(Show Context)
Citation Context ...he reader is referred to the web version of this article.) have been widely used in the polarimetric SAR literature over the last two decades (Ersahin et al., 2004; Ferro-Famil, Pottier, & Lee, 2001; =-=Fukuda & Hirosawa, 1999-=-), and distributed as multi-look processed and publicly available through the polarimetric SAR data processing and educational tool (PolSARpro) by ESA (The Polarimetric SAR Data Processing and Educati... |

19 | Fractional Particle Swarm Optimization in Multi-Dimensional Search Space”,
- Kiranyaz, Ince, et al.
- 2009
(Show Context)
Citation Context ...is still a crucial and challenging task. In this study, RBF network classifier which is optimally designed by the evolutionary search technique, multidimensional particle swarm optimization (MD-PSO) (=-=Kiranyaz, Ince, Yildirim, & Gabbouj, 2010-=-), is employed. RBFs are chosen due to their robustness, faster learning capability compared with other feedforward networks, and superior performance with simpler network architectures. Earlier work ... |

19 | A new classifier for polarimetric SAR images,”
- Zhang, Wu, et al.
- 2009
(Show Context)
Citation Context ...ew wavelet-based texture image segmentation algorithm is successfully applied to unsupervised SAR image segmentation problem. More recently, neural network based approaches (Yang, Wang, & Jiao, 2009; =-=Zhang, Wu, & Wei, 2009-=-; Zhang, Zou, Zhang, & Zhang, 2010) for classification of polarimetric synthetic aperture radar data have been shown to outperform other aforementioned well-known techniques. Compared with other appro... |

18 | Evolutionary artificial neural networks by multi-dimensional particle swarm optimization,”
- Kiranyaz, Ince, et al.
- 2009
(Show Context)
Citation Context ...ery network configuration in the architecture space defined. For PSO-based training, the proposed approach is to apply multi-dimensional particle swarm optimization (MD-PSO) based dynamic clustering (=-=Kiranyaz, Ince, Yildirim, & Gabbouj, 2009-=-) to determine the optimal (with respect to minimizing a given cost function for the input–output mapping) number of Gaussian neurons with their correct parameters (centroids and variances). Afterward... |

17 | Identification of terrain cover using the optimum polarimetric classifier - Kong, Swartz, et al. - 1987 |

13 |
Region-based classification of polarimetric SAR images using Wishart MRF
- Wu, Ji, et al.
- 2008
(Show Context)
Citation Context ...s: (1) classification based on physical scattering mechanisms inherent in data (Pottier & Lee, 2000; van Zyl, 1989), (2) classification based on statistical characteristics of data (Lee et al., 1999; =-=Wu, Ji, Yu, & Su, 2008-=-) and (3) classification based on image processing techniques (Ince, 2010; Tan, Lim, & Ewe, 2007; Ye & Lu, 2002). Additionally, there has been several works using some combinations of the above classi... |

12 | Classification comparisons between dual-pol, compact polarimetric and quadpol SAR imagery - Ainsworth, Kelly, et al. - 2009 |

10 |
Initialization of markov random field clustering of large remote sensing images,”
- Tran, Wehrens, et al.
- 2005
(Show Context)
Citation Context ...ution functions which represent the elements of a 9-dimensional feature vector, consisting of six magnitudes and three angles of a coherency matrix. Markov random field (MRF) clustering based method (=-=Tran, Wehrens, Hoekman, & Buydens, 2005-=-) exploiting the spatial relation between adjacent pixels in polarimetric SAR images has been presented. In (Ye & Lu, 2002), a new wavelet-based texture image segmentation algorithm is successfully ap... |

7 |
Unsupervised classification of polarimetric SAR images by EM algorithm
- Khan, Yang, et al.
- 2007
(Show Context)
Citation Context ... / Expert Systems with Applications 39 (2012) 4710–4717 4711 initialization and its convergence depends on several parameters. Recently, a two-stage unsupervised clustering based on the EM algorithm (=-=Khan, Yang, & Zhang, 2007-=-) is presented for classification of polarimetric SAR images. The EM algorithm estimates parameters of the probability distribution functions which represent the elements of a 9-dimensional feature ve... |

6 |
Ewe, “Image processing in polarimetric SAR images using a hybrid entropy decomposition and maximum likelihood
- Tan, Lim, et al.
- 2007
(Show Context)
Citation Context ...00; van Zyl, 1989), (2) classification based on statistical characteristics of data (Lee et al., 1999; Wu, Ji, Yu, & Su, 2008) and (3) classification based on image processing techniques (Ince, 2010; =-=Tan, Lim, & Ewe, 2007-=-; Ye & Lu, 2002). Additionally, there has been several works using some combinations of the above classification approaches (Lee et al., 1999; Pottier & Lee, 2000). While these approaches to the polar... |

5 | Incorporating texture information into polarimetric radar classification using neural networks
- Ersahin, Scheuchi, et al.
(Show Context)
Citation Context ...ased features combined with the backscattering power (Span), and the gray level co-occurrence matrix (GLCM) based texture features as suggested by the results of previous studies (Clausi & Yue, 2004; =-=Ersahin, Scheuchl, & Cumming, 2004-=-). The performance of the proposed RBF network based classifier is evaluated using the fully polarimetric San Francisco Bay and Flevoland data sets acquired by the NASA/Jet Propulsion Laboratory Airbo... |

4 | Unsupervised classification of scattering mechanisms using radar polarimetry data - Zyl - 1989 |

3 |
Quantitative evaluation of polarimetric classification for agricultural crop mapping
- Chen, Li, et al.
- 2007
(Show Context)
Citation Context ...nd Lee (2009). To compare classification results the same eleven training and testing sets are used with those of the NN-based (Zhang et al., 2009), wavelet-based (Fukuda & Hirosawa, 1999), and ECHO (=-=Chen, Li, Pang, & Tian, 2007-=-) classifiers. In Table 4, the overall accuracies in training and testing areas of the Flevoland dataset for the proposed RBF classifier trained using the BP and MD-PSO algorithms and three state of t... |

2 |
An improved Cloude-Pottier decomposition using H/α/span and complex Wishart classifier for polarimetric
- Fang, Wen, et al.
- 2006
(Show Context)
Citation Context ...verage alpha angle (a), and span calculated by the above noncoherent target decomposition method have been commonly used as polarimetric features of a scatterer in many target classification schemes (=-=Fang, Wen, & Yirong, 2006-=-; Lee et al., 1999). 3. Feature extraction The proposed feature extraction process utilizes the complete covariance matrix information, the gray level co-occurrence matrix (GLCM) based texture feature... |

2 |
Unsupervised classification of polarimetric SAR image with dynamic clustering: An image processing approach
- Ince
(Show Context)
Citation Context ...er & Lee, 2000; van Zyl, 1989), (2) classification based on statistical characteristics of data (Lee et al., 1999; Wu, Ji, Yu, & Su, 2008) and (3) classification based on image processing techniques (=-=Ince, 2010-=-; Tan, Lim, & Ewe, 2007; Ye & Lu, 2002). Additionally, there has been several works using some combinations of the above classification approaches (Lee et al., 1999; Pottier & Lee, 2000). While these ... |

1 | Polarimetric SAR image classification using a radial basis function neural network
- Ince
- 2010
(Show Context)
Citation Context ...er & Lee, 2000; van Zyl, 1989), (2) classification based on statistical characteristics of data (Lee et al., 1999; Wu, Ji, Yu, & Su, 2008) and (3) classification based on image processing techniques (=-=Ince, 2010-=-; Tan, Lim, & Ewe, 2007; Ye & Lu, 2002). Additionally, there has been several works using some combinations of the above classification approaches (Lee et al., 1999; Pottier & Lee, 2000). While these ... |

1 | Cheng-Chang (2002). Wavelet-based unsupervised SAR image segmentation using hidden markov tree models - Ye, Lu |