# NCT-CRC-HE: Not All Histopathological Datasets Are Equally Useful

arXiv: https://arxiv.org/abs/2409.11546 · HTML: https://mlsanity.com/paper/2409.11546/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 15 of its 59 references that are in the map.
- First posted on arXiv 17 Sep 2024.

## Builds on (most cited first, 15 of 15)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://mlsanity.com/paper/1905.11946/) (2019)
- [Whole Slide Images based Cancer Survival Prediction using Attention Guided Deep Multiple Instance Learning Networks](https://mlsanity.com/paper/2009.11169/) (2020)
- [Quantifying the effects of data augmentation and stain color normalization in convolutional neural networks for computational pathology](https://mlsanity.com/paper/1902.06543/) (2019)
- [Patch-based Convolutional Neural Network for Whole Slide Tissue Image Classification](https://mlsanity.com/paper/1504.07947/) (2015)
- [Benchmarking Self-Supervised Learning on Diverse Pathology Datasets](https://mlsanity.com/paper/2212.04690/) (2022)
- [Towards a Visual-Language Foundation Model for Computational Pathology](https://mlsanity.com/paper/2307.12914/) (2023)
- [Deep learning-based survival prediction for multiple cancer types using histopathology images](https://mlsanity.com/paper/1912.07354/) (2019)
- [UniToPatho, a labeled histopathological dataset for colorectal polyps classification and adenoma dysplasia grading](https://mlsanity.com/paper/2101.09991/) (2021)
- [SlideGraph+: Whole Slide Image Level Graphs to Predict HER2Status in Breast Cancer](https://mlsanity.com/paper/2110.06042/) (2021)
- [StainNet: a fast and robust stain normalization network](https://mlsanity.com/paper/2012.12535/) (2020)
- [DiagSet: a dataset for prostate cancer histopathological image classification](https://mlsanity.com/paper/2105.04014/) (2021)
- [All You Need is Color: Image based Spatial Gene Expression Prediction using Neural Stain Learning](https://mlsanity.com/paper/2108.10446/) (2021)
- [Large-scale Gastric Cancer Screening and Localization Using Multi-task Deep Neural Network](https://mlsanity.com/paper/1910.03729/) (2019)
- [Histopathological Image Classification with Cell Morphology Aware Deep Neural Networks](https://mlsanity.com/paper/2407.08625/) (2024)

## Built on by (most cited first, 5 of 5)

- [SAG-ViT: A Scale-Aware, High-Fidelity Patching Approach with Graph Attention for Vision Transformers](https://mlsanity.com/paper/2411.09420/) (2024)
- [Colorectal Cancer Tumor Grade Segmentation in Digital Histopathology Images: From Giga to Mini Challenge](https://mlsanity.com/paper/2507.04681/) (2025)
- [Intuitions of Machine Learning Researchers about Transfer Learning for Medical Image Classification](https://mlsanity.com/paper/2510.00902/) (2025)
- [A Dataset is Worth 1 MB](https://mlsanity.com/paper/2602.23358/) (2026)
- [CARA: Collision-Aware Resolution Adaptation for Multiresolution Hash Encoding Based Image Fitting](https://mlsanity.com/paper/2609.18554/) (2026)

## Abstract (arXiv)

Numerous deep learning-based solutions have been proposed for histopathological image analysis over the past years. While they usually demonstrate exceptionally high accuracy, one key question is whether their precision might be affected by low-level image properties not related to histopathology but caused by microscopy image handling and pre-processing. In this paper, we analyze a popular NCT-CRC-HE-100K colorectal cancer dataset used in numerous prior works and show that both this dataset and the obtained results may be affected by data-specific biases. The most prominent revealed dataset issues are inappropriate color normalization, severe JPEG artifacts inconsistent between different classes, and completely corrupted tissue samples resulting from incorrect image dynamic range handling. We show that even the simplest model using only 3 features per image (red, green and blue color intensities) can demonstrate over 50% accuracy on this 9-class dataset, while using color histogram not explicitly capturing cell morphology features yields over 82% accuracy. Moreover, we show that a basic EfficientNet-B0 ImageNet pretrained model can achieve over 97.7% accuracy on this dataset, outperforming all previously proposed solutions developed for this task, including dedicated foundation histopathological models and large cell morphology-aware neural networks. The NCT-CRC-HE dataset is publicly available and can be freely used to replicate the presented results. The codes and pre-trained models used in this paper are available at https://github.com/gmalivenko/NCT-CRC-HE-experiments