Large-Margin Representation Learning for Texture Classification

TL;DR

This paper introduces a large-margin representation learning method combining convolutional layers and metric learning, enabling effective texture classification on small datasets with lower computational cost and faster convergence.

Contribution

It proposes a novel loss function and training approach that improves texture classification by reducing parameters and computational cost compared to traditional CNNs.

Findings

Achieves competitive accuracy on texture and histopathologic datasets.

Reduces training cost and converges faster than CNNs.

Effective with small datasets due to fewer parameters.

Abstract

This paper presents a novel approach combining convolutional layers (CLs) and large-margin metric learning for training supervised models on small datasets for texture classification. The core of such an approach is a loss function that computes the distances between instances of interest and support vectors. The objective is to update the weights of CLs iteratively to learn a representation with a large margin between classes. Each iteration results in a large-margin discriminant model represented by support vectors based on such a representation. The advantage of the proposed approach w.r.t. convolutional neural networks (CNNs) is two-fold. First, it allows representation learning with a small amount of data due to the reduced number of parameters compared to an equivalent CNN. Second, it has a low training cost since the backpropagation considers only support vectors. The experimental results on texture and histopathologic image datasets have shown that the proposed approach achieves competitive accuracy with lower computational cost and faster convergence when compared to equivalent CNNs.