Optimizing Gabor Texture Features for Materials Recognition by Convolutional Neural Networks

In this paper, we present a novel technique that allows for customized Gabor texture features by leveraging deep learning neural networks. Our method involves using Convolutional Neural Network to refactor traditional, hand-designed filters on specific datasets. The refactored can be used in an off-the-shelf manner with the same computational cost but significantly improved accuracy material recognition. We demonstrate effectiveness of our approach reporting gain discriminatio different is particularly appealing situations where use entire CNN would inadequate, such as analyzing non-square images or performing segmentation tasks. Overall, provides powerful tool improving recognition tasks while retaining advantages handcrafted filters.