Independent Component Analysis Based Incoherent Target Decompositions for Polarimetric SAR Data ‐ Practical Aspects

Vasile, Gabriel

doi:10.1109/igarss.2018.8518738

Cited by 7 publications

(5 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The proposed ICA was able to retrieve nonorthogonal scatterer types [10,11]. By applying the MMSE filter [12] on each of the ICA derived rotation invariant scattering vectors, we have shown in [13] that spatial resolution can be better preserved with respect to the conventional Pol-SAR boxcar speckle filter.…”

Section: Introductionmentioning

confidence: 89%

See 1 more Smart Citation

On ICA Based ICTD Classification of Polsar Data

Vasile

2019

IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium

Self Cite

View full text Add to dashboard Cite

The Independent Component Analysis (ICA) has been recently introduced as a reliable alternative to identify canonical scattering mechanisms within PolSAR images. This paper addresses an important aspect for applying such methods on real data, namely statistical classification with ICA. A novel algorithm is proposed by adjusting the iterative segmentation from [1, 2] to the particular nature of the Touzi's polarimetric decomposition [3]. This algorithm is tested using P-band airborne PolSAR data acquired for the ESA campaign TropiSAR campaign.

show abstract

Section: Introductionmentioning

confidence: 89%

“…The only restriction would be that the use of the ICA scattering mechanisms are not directly rotational invariant. As presented in [13], we propose once again to use the rotation invariant scattering vectors of the following form [3]:…”

Section: Ica-ictd Statistical Classificationmentioning

confidence: 99%

On ICA Based ICTD Classification of Polsar Data

Vasile

2019

IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the local approach, by applying the MMSE filter [7] on each of the ICA derived rotation invariant scattering vectors, we have shown in [8] that spatial resolution can be better preserved with respect to the conventional PolSAR boxcar speckle filter.…”

Section: Incoherent Target Decomposition Techniques For Polsar Datamentioning

confidence: 99%

Comparative Analysis of ICA Based Incoherent Target Decompositions for Polsar Data

Vasile

Congedo

2022

IGARSS 2022 - 2022 IEEE International Geoscience and Remote Sensing Symposium

Self Cite

View full text Add to dashboard Cite

The Independent Component Analysis (ICA) has been recently introduced as a reliable alternative to identify canonical scattering mechanisms within PolSAR images. This paper addresses an overview of the most important aspects for applying such methods on real data. A new geometric classification algorithm is proposed by combining the polar decomposition and by adjusting the conventional k-means mean and distance with their counterpart on the Riemannian manifold. This algorithm is tested using P-band airborne PolSAR data acquired for the ESA campaign TropiSAR campaign.

show abstract

“…The feature dimension of SAR images can be effectively reduced by using feature extraction, and its quality directly affects the classification of subsequent classifiers. The generally used feature extraction methods include principal component analysis (PCA) [11], [12], linear decision analysis (LDA) [13], independent component analysis (ICA) [14], fast fourier transform (FFT) [15] and ratio detector (RD) [16], etc. It is difficult to satisfactorily classify SAR image targets by using general classifiers directly.…”

Section: Introductionmentioning

confidence: 99%

MdpCaps-Csl for SAR Image Target Recognition With Limited Labeled Training Data

Hou

et al. 2020

IEEE Access

View full text Add to dashboard Cite

Although convolutional neural networks (CNN) have shown excellent performance in many image recognition tasks, it commonly requires a lot of labeled data, and the recognition effect is frequently unsatisfied due to the limited labeled training data. In recent years, capsule network (CapsNet) has been shown to achieve a high recognition accuracy with a small group of training samples. In this study, a class separable loss based on cosine similarity is suggested to enhance the distinguishability of the extracted network. It is added as a regularization term to the original loss function to train the network, narrowing the intra-class difference and increasing the inter-class difference in each iteration. Meanwhile, a multidimensional parallel capsule module is established to obtain robust features and spatial relationships from the original images. Feature maps from convolution of different levels are extracted as the input of this module. Structural features derived from low-level convolution and semantic features derived from highlevel convolution are used for low-dimensional capsule coding and high-dimensional capsule coding, respectively. In our experiment, the general moving and stationary target acquisition and recognition (MSTAR) database is used. We find that the accuracy of the multi-dimensional parallel capsule network with class separable loss (MdpCaps-Csl) is 99.79% using all training samples, which is higher than most current recognition methods. More importantly, the accuracy is up to 97.73% even if only 10% training samples are applied, indicating MdpCaps-Csl can make excellent performance upon limited training samples. INDEX TERMS capsule network, class separable loss, image target recognition, limited labeled training data, multi-dimensional parallel capsule module, synthetic aperture radar.

show abstract

Independent Component Analysis Based Incoherent Target Decompositions for Polarimetric SAR Data ‐ Practical Aspects

Cited by 7 publications

References 11 publications

On ICA Based ICTD Classification of Polsar Data

On ICA Based ICTD Classification of Polsar Data

Comparative Analysis of ICA Based Incoherent Target Decompositions for Polsar Data

MdpCaps-Csl for SAR Image Target Recognition With Limited Labeled Training Data

Contact Info

Product

Resources

About