Skip to Main content Skip to Navigation
Theses

Kernel-based learning on hierarchical image representations : applications to remote sensing data classification

Abstract : Hierarchical image representations have been widely used in the image classification context. Such representations are capable of modeling the content of an image through a tree structure. In this thesis, we investigate kernel-based strategies that make possible taking input data in a structured form and capturing the topological patterns inside each structure through designing structured kernels. We develop a structured kernel dedicated to unordered tree and path (sequence of nodes) structures equipped with numerical features, called Bag of Subpaths Kernel (BoSK). It is formed by summing up kernels computed on subpaths (a bag of all paths and single nodes) between two bags. The direct computation of BoSK yields a quadratic complexity w.r.t. both structure size (number of nodes) and amount of data (training size). We also propose a scalable version of BoSK (SBoSK for short), using Random Fourier Features technique to map the structured data in a randomized finite-dimensional Euclidean space, where inner product of the transformed feature vector approximates BoSK. It brings down the complexity from quadratic to linear w.r.t. structure size and amount of data, making the kernel compliant with the large-scale machine-learning context. Thanks to (S)BoSK, we are able to learn from cross-scale patterns in hierarchical image representations. (S)BoSK operates on paths, thus allowing modeling the context of a pixel (leaf of the hierarchical representation) through its ancestor regions at multiple scales. Such a model is used within pixel-based image classification. (S)BoSK also works on trees, making the kernel able to capture the composition of an object (top of the hierarchical representation) and the topological relationships among its subparts. This strategy allows tile/sub-image classification. Further relying on (S)BoSK, we introduce a novel multi-source classification approach that performs classification directly from a hierarchical image representation built from two images of the same scene taken at different resolutions, possibly with different modalities. Evaluations on several publicly available remote sensing datasets illustrate the superiority of (S)BoSK compared to state-of-the-art methods in terms of classification accuracy, and experiments on an urban classification task show the effectiveness of proposed multi-source classification approach.
Complete list of metadatas

Cited literature [221 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-01717563
Contributor : Abes Star :  Contact
Submitted on : Monday, February 26, 2018 - 2:59:14 PM
Last modification on : Thursday, August 27, 2020 - 4:07:29 PM
Long-term archiving on: : Monday, May 28, 2018 - 4:22:55 PM

File

2017theseCuiY.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-01717563, version 1

Citation

Yanwei Cui. Kernel-based learning on hierarchical image representations : applications to remote sensing data classification. Computer Vision and Pattern Recognition [cs.CV]. Université de Bretagne Sud, 2017. English. ⟨NNT : 2017LORIS448⟩. ⟨tel-01717563⟩

Share

Metrics

Record views

573

Files downloads

113