Hadamard-Gabor Filter: A Rotation Invariant Feature Extraction Algorithm

Gabor filter is a well-known feature extraction algorithm in image processing with many applications in texture segmentation, fractal dimension management, document analysis, edge detection, retina identification, image coding and image representation. The Gabor filter has received considerable attention mainly because the human visual system is similar to this filter. In this paper we propose Hadamard-Gabor filter a rotation-invariant feature extraction algorithm. In our new Gabor filter we apply Walsh-Hadamard and Johnson-Lindenstrauss transforms on the features of an image extracted using Gabor filter. The Walsh-Hadamard transform makes Gaber filter rotation-invariant and the Johnson-Lindenstrauss transform reduces the dimension of extracted features to a lower (target) dimension. Many image processing algorithms use feature extraction algorithms as one of their basic building blocks. Using Hadamard-Gabor filter instead of Gabor filter makes them potentially faster as they deal with features of lower dimensions.