Filtering an image by Gabor wavelet is one of the widely used methods for feature extraction. Convolutioning an image with Gabor filters generates transformed images. Below image shows 200 Gabor filters that can extract features from images almost as similar as a human visual system does.
Description of the convolution of an image with a few Gabor filters.
After the implementation, you can choose a few of Gabor filters that effectively extract the feature of images. It’s more or less like PCA(Principal Component Analysis)