Image Classification Using Singular Value Decomposition and Optimization

TL;DR

The results partially validate the assumption that dominant features, such as fur color, can be effectively captured through low-rank approximations and suggest that additional features or methods may be required for more robust classification, highlighting the trade-off between simplicity and performance in resource-constrained environments.

Abstract

This study investigates the applicability of Singular Value Decomposition for the image classification of specific breeds of cats and dogs using fur color as the primary identifying feature. Sequential Quadratic Programming (SQP) is employed to construct optimally weighted templates. The proposed method achieves 69% accuracy using the Frobenius norm at rank 10. The results partially validate the assumption that dominant features, such as fur color, can be effectively captured through low-rank approximations. However, the accuracy suggests that additional features or methods may be required for more robust classification, highlighting the trade-off between simplicity and performance in resource-constrained environments.

Figures (7)

• Figure 1: Persian Cat and Boxer Dog templates for uniform and optimal weights.
• Figure 2: Comparison of norms: 1 norm, 2 norm, $\infty$ norm, and Frobenius norm across different ranks for the training set.
• Figure 3: Low-rank representations of a test Persian cat and a test Boxer dog for each norm at the best rank.
• Figure 4: Classification metrics: TP Recall, FP Recall, and Accuracy across norms at respective optimal ranks for the training set.
• Figure 5: Test set classification of Boxer dogs vs. Persian cats, frobenius norm, rank = 10.