Document Type
Technical Report
Publication Date
10-1-2004
Technical Report Number
TR2005-545
Abstract
In this paper, we propose a new method for automatically determining image orientations. This method is based on a set of natural image statistics collected from a multi-scale multi-orientation image decomposition (e.g., wavelets). From these statistics, a two-stage hierarchal classification with multiple binary SVM classifiers is employed to de- termine image orientation. The proposed method is evaluated and compared to existing methods with experiments performed on 18040 natural images, where it showed promising performance.
Dartmouth Digital Commons Citation
Lyu, Siwei, "Automatic Image Orientation Determination with Natural Image Statistics" (2004). Computer Science Technical Report TR2005-545. https://digitalcommons.dartmouth.edu/cs_tr/273