breastmnist (MedMNIST)

Source: breastmnist
Task: binary-class
Resolutions: 224x224
License: CC BY 4.0

Description

The BreastMNIST is based on a dataset of 780 breast ultrasound images. It is categorized into 3 classes: normal, benign, and malignant. As we use low-resolution images, we simplify the task into binary classification by combining normal and benign as positive and classifying them against malignant as negative. We split the source dataset with a ratio of 7:1:2 into training, validation and test set. The source images of 1×500×500 are resized into 1×28×28.

Config naming convention

{split}-{class}-{res}

split : train | val | test
class : all   | <sanitized class name>
res   : res28 | res64 | res128 | res224

Loading examples

from datasets import load_dataset

# All training images at 224px
ds = load_dataset('.../breastmnist', 'train-all-res224', split='train')

# Only 'malignant' class, training split
ds = load_dataset('.../breastmnist', 'train-malignant-res224', split='train')

Class labels

• 0 — malignant (config key: malignant)
• 1 — normal, benign (config key: normal_benign)

Class distribution

224x224

train (N=546, IR=2.71x)

Class	Config key	Count	Share
malignant	malignant	147	26.9%
normal, benign	normal_benign	399	73.1%

val (N=78, IR=2.71x)

Class	Config key	Count	Share
malignant	malignant	21	26.9%
normal, benign	normal_benign	57	73.1%

test (N=156, IR=2.71x)

Class	Config key	Count	Share
malignant	malignant	42	26.9%
normal, benign	normal_benign	114	73.1%

Citation

@article{medmnistv2,
  title={MedMNIST v2 - A large-scale lightweight benchmark for 2D and 3D biomedical image classification},
  author={Yang, Jiancheng and Shi, Rui and Wei, Donglai and Liu, Zequan
          and Zhao, Lin and Ke, Bilian and Pfister, Hanspeter and Ni, Bingbing},
  journal={Scientific Data},
  year={2023}
}