Fisher Vectors for PolSAR Image Classification

Redolfi, Javier A.; Sánchez, Jorge M. Balestena; Flesia, Ana Georgina

doi:10.1109/lgrs.2017.2750800

Cited by 8 publications

(4 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, divergent polarization decomposition features can be computed with given specific scattering basis sets [13][14][15] or different eigenvalue decomposition [16,17]. Second, features are extracted with machine-learning and computer-vision methods as in [18][19][20]. More recently, with the idea of learning theory, Refs.…”

Section: Related Workmentioning

confidence: 99%

Structure Label Matrix Completion for PolSAR Image Classification

Hou

Wen

et al. 2020

Remote Sensing

View full text Add to dashboard Cite

Terrain classification is a hot topic in polarimetric synthetic aperture radar (PolSAR) image interpretation that aims at assigning a label to every pixel and forms a label matrix for a PolSAR image. From the perspective of human interpretation, classification is not in terms of pixels, but decomposed perceptual groups and structures. Therefore, a new perspective of label matrix completion is proposed to treat the classification task as a label matrix inversion process. Firstly, a matrix completion framework is built to avoid processing the large-scale PolSAR image data for traditional feature and classifier learning. Secondly, a light network is used to obtain the known labels for the complete task by uniform down-sampling of the entire image that aims to keep the shape of regions in a PolSAR image and reduce the computational complexity. Finally, the zeroth-and first-order label information is proposed as the prior distribution for label matrix completion to keep the structure and texture. The experiments are tested on real PolSAR images, which demonstrates that the proposed method can realize excellent classification results with less computation time and labeled pixels. Classification results with different down-sampling rates in the light network also prove the robustness of this method.

show abstract

Section: Related Workmentioning

confidence: 99%

Structure Label Matrix Completion for PolSAR Image Classification

Hou

Wen

et al. 2020

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…In this paper we propose two modifications to the work presented in [Redolfi et al, 2017]. The first one is with respect to the selection of the pdf of equation 5.…”

Section: Proposed Extensionmentioning

confidence: 99%

“…The main contributions of this work are the extension of a previously presented method [Redolfi et al, 2017], adding the comparison of different probability distributions functions for the eFVs encoding and different similarity measures between the eFVs to generate a raw classification of pixels. The hypotheses of this work are that although the theoretical distribution of the data follows a Wishart distribution, with Gaussian distributions competitive results can be obtained and also that using other types of similarity measures the results in the classification can be improved.…”

Section: Introductionmentioning

confidence: 99%

Polsar Image Classification Using Different Codifications Based on Fisher Vectors

Redolfi¹,

Araguás²,

Flesia

2020

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

Abstract. A PolSAR is an active sensing device capable of providing images that are robust against variations of weather and atmosphere conditions, irrespective of the time of the day they were acquired. For an efficient use of these images it is necessary to have algorithms capable of classifying these images to generate maps with their content automatically. This paper presents the extension of a PolSAR image classification method based on exponential Fisher Vectors, a Potts smoothing model and different similarity measures. With the proposed extension, improvements in classification with respect to the base method are achieved. Future work consists in extending the codification so as not to have to discard the imaginary part of the data.

show abstract

“…All the above methods incorporate the prior knowledge to capture SAR image features, and these features are poor generalization. Afterwards, with the purpose of achieving stronger features, there has been increasing interests in the middlelevel feature learning methods based on machine learning techniques, i.e., sparse representation [12], fisher vector (FV) [13], superpixel-level FV [14] and multi-scale local fisher patterns [15]. Although these feature descriptors characterize SAR images better than the low-level features, they require lots of prior knowledge for feature extraction and are not suitable for HR SAR image processing under amounts of SAR data.…”

mentioning

confidence: 99%

A Wavelet-Driven Subspace Basis Learning Network for High-Resolution Synthetic Aperture Radar Image Classification

Zhai

et al. 2023

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Fisher Vectors for PolSAR Image Classification

Cited by 8 publications

References 22 publications

Structure Label Matrix Completion for PolSAR Image Classification

Structure Label Matrix Completion for PolSAR Image Classification

Polsar Image Classification Using Different Codifications Based on Fisher Vectors

A Wavelet-Driven Subspace Basis Learning Network for High-Resolution Synthetic Aperture Radar Image Classification

Contact Info

Product

Resources

About