Learning 2D Gabor Filters by Infinite Kernel Learning Regression

TL;DR

This paper introduces a novel method for learning dataset-specific 2D Gabor filters using infinite kernel learning regression, resulting in sparse, effective filters that enhance face recognition accuracy.

Contribution

It proposes a new formulation for image representation with Gabor functions based on infinite kernel learning, enabling the learning of dataset-specific Gabor filters.

Findings

Sparse Gabor filter representations improve recognition accuracy.

Support vector expansion effectively models images with Gabor functions.

Learning dataset-specific filters enhances face recognition performance.

Abstract

Gabor functions have wide-spread applications in image processing and computer vision. In this paper, we prove that 2D Gabor functions are translation-invariant positive-definite kernels and propose a novel formulation for the problem of image representation with Gabor functions based on infinite kernel learning regression. Using this formulation, we obtain a support vector expansion of an image based on a mixture of Gabor functions. The problem with this representation is that all Gabor functions are present at all support vector pixels. Applying LASSO to this support vector expansion, we obtain a sparse representation in which each Gabor function is positioned at a very small set of pixels. As an application, we introduce a method for learning a dataset-specific set of Gabor filters that can be used subsequently for feature extraction. Our experiments show that use of the learned Gabor filters improves the recognition accuracy of a recently introduced face recognition algorithm.