Understanding the impact of image and input resolution on deep digital pathology patch classifiers
This work addresses annotation efficiency in digital pathology, where expert labels are scarce, offering incremental improvements for cancer diagnosis.
The study investigated how image and input resolution affect deep learning patch classifiers in digital pathology, finding that manipulating these resolutions improves classification accuracy, with a model trained on less than 1% of data matching the performance of one trained on full data in the PCam dataset.
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.