A Theoretical Analysis of Deep Neural Networks for Texture Classification

TL;DR

This paper provides a theoretical analysis of deep neural networks for texture classification, deriving bounds on VC dimension, and exploring the impact of intrinsic data dimensionality on network performance.

Contribution

It introduces the first upper bounds on VC dimension for CNNs, Dropout, and Dropconnect networks, and analyzes how texture dataset complexity affects neural network shattering.

Findings

Hand-crafted features reduce excess error rate.

Texture datasets are inherently higher dimensional.

Relative Contrast diminishes with increasing data dimensionality.

Abstract

We investigate the use of Deep Neural Networks for the classification of image datasets where texture features are important for generating class-conditional discriminative representations. To this end, we first derive the size of the feature space for some standard textural features extracted from the input dataset and then use the theory of Vapnik-Chervonenkis dimension to show that hand-crafted feature extraction creates low-dimensional representations which help in reducing the overall excess error rate. As a corollary to this analysis, we derive for the first time upper bounds on the VC dimension of Convolutional Neural Network as well as Dropout and Dropconnect networks and the relation between excess error rate of Dropout and Dropconnect networks. The concept of intrinsic dimension is used to validate the intuition that texture-based datasets are inherently higher dimensional as compared to handwritten digits or other object recognition datasets and hence more difficult to be shattered by neural networks. We then derive the mean distance from the centroid to the nearest and farthest sampling points in an n-dimensional manifold and show that the Relative Contrast of the sample data vanishes as dimensionality of the underlying vector space tends to infinity.