Datasets:

xtxx
/

HepatoBench

Tasks:

Image Classification

Modalities:

Size:

License:

Dataset card Data Studio Files Files and versions

You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

HepatoBench Dataset

HepatoBench is a seven-class liver histopathology patch benchmark curated for evaluating pathology foundation models.

Dataset page: https://huggingface.co/datasets/xtxx/HepatoBench
Task: Patch-level image classification (7 classes)
Release: meta Information (dataset_meta.json)

Overview

HepatoBench contains 87,301 liver tissue image patches from seven categories:

Class	Description	# Images
TUM	Tumor	13,200
FIB	Fibrosis	26,364
INF	Inflammation	4,776
NEC	Necrosis	22,023
NOR	Normal Tissue	13,200
REA	Biliary Reaction	1,540
STE	Steroidal Degeneration	6,198

License

The dataset is licensed under CC BY-NC 4.0 and is for research use only.

Contacts

Email: lingxt23@mails.tsinghua.edu.cn

Downloads last month: -

Collection including xtxx/HepatoBench

HepatoSource

2 items • Updated Jan 6