Intelligent Multi-View Test Time Augmentation

TL;DR

This paper presents an adaptive test time augmentation method that intelligently selects optimal image augmentations based on uncertainty metrics, improving classification accuracy under viewpoint variations.

Contribution

It introduces a novel two-stage uncertainty-based selection process for TTA, enhancing robustness and accuracy over traditional uniform augmentation methods.

Findings

Achieved an average accuracy increase of 1.73% across multiple datasets.

Validated effectiveness across various neural network architectures.

Demonstrated improved robustness against viewpoint variations.

Abstract

In this study, we introduce an intelligent Test Time Augmentation (TTA) algorithm designed to enhance the robustness and accuracy of image classification models against viewpoint variations. Unlike traditional TTA methods that indiscriminately apply augmentations, our approach intelligently selects optimal augmentations based on predictive uncertainty metrics. This selection is achieved via a two-stage process: the first stage identifies the optimal augmentation for each class by evaluating uncertainty levels, while the second stage implements an uncertainty threshold to determine when applying TTA would be advantageous. This methodological advancement ensures that augmentations contribute to classification more effectively than a uniform application across the dataset. Experimental validation across several datasets and neural network architectures validates our approach, yielding an average accuracy improvement of 1.73% over methods that use single-view images. This research underscores the potential of adaptive, uncertainty-aware TTA in improving the robustness of image classification in the presence of viewpoint variations, paving the way for further exploration into intelligent augmentation strategies.