Extended Dependency Tree-HMM for Non-Rectangular Sub-Images Modeling

Mohamed El Yazid Boudaren (1) , Abdel Belaïd (2)

1 Applied Mathematics Laboratory [EMP - Algiers]
2 READ - READ

This work is motivated by the need of evaluating the likelihood probability on sub-images of not necessarily rectangular shape in some frameworks. For this purpose, we propose an alternative of Dependency Tree- HMM that allows the four traditional interactions between neighboring pixels instead of just two. To demonstrate the accuracy of the proposed model, we provide some classification results performed on high resolution aerial images.

Domaines: Traitement du signal et de l'image [eess.SP]; Traitement du signal et de l'image [eess.SP]

Mohamed El Yazid Boudaren, Abdel Belaïd. Extended Dependency Tree-HMM for Non-Rectangular Sub-Images Modeling. The Fifth International Workshop on Applied Probability - IWAP 2010, Universidad Carlos III de Madrid, Jul 2010, Madrid, Spain. pp.4. ⟨inria-00579675⟩