Transformation Invariant Cancerous Tissue Classification Using Spatially Transformed DenseNet

TL;DR

This paper presents a spatially transformed DenseNet model that enhances cancer tissue classification accuracy while achieving transformation invariance, offering a simpler alternative to existing invariant models.

Contribution

The paper introduces a novel spatially transformed DenseNet architecture that improves accuracy and invariance in cancer tissue classification with reduced complexity.

Findings

Increased classification accuracy over baseline DenseNet.

Achieved transformation invariance in tissue classification.

Simpler architecture compared to existing invariant models.

Abstract

In this work, we introduce a spatially transformed DenseNet architecture for transformation invariant classification of cancer tissue. Our architecture increases the accuracy of the base DenseNet architecture while adding the ability to operate in a transformation invariant way while simultaneously being simpler than other models that try to provide some form of invariance.