Abstract

A fast and accurate texture recognition system is presented. The new approach consists in extracting locally and globally invariant representations. The locally invariant representation is built on a multi-resolution convolutional net- work with a local pooling operator to improve robustness to local orientation and scale changes. This representation is mapped into a globally invariant descriptor using multifractal analysis. We propose a new multifractal descriptor that cap- tures rich texture information and is mathematically invariant to various complex transformations. In addition, two more techniques are presented to further im- prove the robustness of our system. The first technique consists in combining the generative PCA classi fier with multiclass SVMs. The second technique consists of two simple strategies to boost classi fication results by synthetically augment- ing the training set. Experiments show that the proposed solution outperforms existing methods on three challenging public benchmark datasets, while being computationally efficient.