Linear Analysis Metric Learning using Siamese Neural Networks

Abin Jose, Qi Mei

Submitted to IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP 2022),
Conference homepage


We propose a method for learning the Linear Discriminant Analysis (LDA) using a Siamese Neural Network (SNN) architecture. The novelty of our work is that we learn the LDA projection matrix between the final fully-connected layers of an SNN. An SNN architecture is used since the proposed loss maximizes the Kullback-Leibler divergence between the feature distributions from the two branches of an SNN. The network learns an optimized feature space having inherent properties pertaining to the learning of LDA. The learned image descriptors are a) low-dimensional, b) have small intra-class variance c) large inter-class variance, and d) is able to classify the classes with linear decision hyperplanes. The proposed method has the advantage that the LDA learning happens end-to-end. We measured the classification accuracy in three datasets such as MNIST, CIFAR-10, and STL-10 and compared the performance with other state-of-the-art methods. We also measured the KL divergence between the class pairs and visualized the projection of feature vectors along the learned discriminant directions.

Supplementary Material

Network architecture and parameter tuning

NOTICE FOR IEEE PUBLICATIONS: © IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works, must be obtained from the IEEE. Contact: Manager, Copyrights and Permissions / IEEE Service Center / 445 Hoes Lane / P.O. Box 1331 / Piscataway, NJ 08855-1331, USA. Telephone: + Intl. 908-562-3966.