Tile-Level Annotation of Satellite Images Using Multi-Level Max-Margin Discriminative Random Field

Fan Hu,Wen Yang,Jiayu Chen,Hong Sun

doi:10.3390/rs5052275

Abstract

This paper proposes a multi-level max-margin discriminative analysis (M3DA) framework, which takes both coarse and fine semantics into consideration, for the annotation of high-resolution satellite images. In order to generate more discriminative topic-level features, the M3DA uses the maximum entropy discrimination latent Dirichlet Allocation (MedLDA) model. Moreover, for improving the spatial coherence of visual words neglected by M3DA, conditional random field (CRF) is employed to optimize the soft label field composed of multiple label posteriors. The framework of M3DA enables one to combine word-level features (generated by support vector machines) and topic-level features (generated by MedLDA) via the bag-of-words representation. The experimental results on high-resolution satellite images have demonstrated that, using the proposed method can not only obtain suitable semantic interpretation, but also improve the annotation performance by taking into account the multi-level semantics and the contextual information.

Highlights

Nowadays the information extraction and intelligent interpretation of high-resolution satellite images are frontier technologies in the remote sensing field
There have been ever-growing interests in image annotation by using topic models, such as Probabilistic Latent Semantic Analysis (PLSA) [1,2], Latent Dirichlet Allocation (LDA) [3,4], which can map from low-level physical features to high-level semantic concepts, and essentially reduce the dimensionality of features
We introduce the conditional random field (CRF) model over the label inference in soft label fields generated by the multi-level max-margin discriminative topic model

Summary

Introduction

Nowadays the information extraction and intelligent interpretation of high-resolution satellite images are frontier technologies in the remote sensing field. There have been ever-growing interests in image annotation by using topic models, such as Probabilistic Latent Semantic Analysis (PLSA) [1,2], Latent Dirichlet Allocation (LDA) [3,4], which can map from low-level physical features to high-level semantic concepts, and essentially reduce the dimensionality of features. These generative probabilistic models were originally developed for text document modeling, which can generate an infinite sequence of samples according to the distribution of latent topics. Local features extracted from patches are transformed by vector quantization into “visual words”, and each tile is represented as a collection of words

Methods

Results

Discussion

Conclusion