Understanding the impact of image and input resolution on deep digital pathology patch classifiers

TL;DR

This paper investigates how varying image and input resolutions affect deep learning-based patch classification in digital pathology, demonstrating that resolution adjustments can improve performance especially with limited annotations.

Contribution

It provides a systematic analysis of resolution effects in digital pathology classification, showing that resolution manipulation can compensate for scarce annotations and maintain high accuracy.

Findings

Higher image and input resolutions improve classification accuracy.

Resolution tuning allows models trained on less data to match full-data performance.

Optimal resolution strategies can reduce annotation costs significantly.

Abstract

We consider annotation efficient learning in Digital Pathology (DP), where expert annotations are expensive and thus scarce. We explore the impact of image and input resolution on DP patch classification performance. We use two cancer patch classification datasets PCam and CRC, to validate the results of our study. Our experiments show that patch classification performance can be improved by manipulating both the image and input resolution in annotation-scarce and annotation-rich environments. We show a positive correlation between the image and input resolution and the patch classification accuracy on both datasets. By exploiting the image and input resolution, our final model trained on < 1% of data performs equally well compared to the model trained on 100% of data in the original image resolution on the PCam dataset.