SAR Image Labeling with Hierarchical Markov Aspect Models
Dengxin Dai, Yang Wen and William Triggs
IEEE Transactions on Geoscience and Remote Sensing Volume In press, 2009.

Abstract

Scene segmentation and semantic labeling are important problems in SAR image interpretation. This paper proposes an efficient SAR imagery labeling method based on aspect model which can be learnt from keywords-labeled training data directly. Furthermore, a novel hierarchical Markov aspect model (HMAM) is presented by building aspect model on quadtree. HMAM outperform both aspect model and hierarchical MRFs due to their complementary as aspect model use global relevance estimates while quadtree can further explore image context and multi-scale cues. The experimental results on TerraSAR-X dataset show that our labeling method is effective and efficient, and demonstrate that HMAM improve labeling performance significantly with only a modest increase in learning and inference complexity than aspect model.

PDF - Requires Adobe Acrobat Reader or other PDF viewer.

[Edit]