The LDCH dataset contains 917 cervical histopathological whole-slide images: Cervicitis (156), CINI (305), CINII (296), CINIII (162), Cancer (25). Among them, 286 images were released with local annotations to mark out several lesion areas. The areas can be classified into 9 categories: CINI, CINII, CINIII, CINII involve gland, CINIII involve gland, gland, gland squamous metaplasia, koilocyte and squamous epithelium.

The images were collected by Landing Artificial Intelligence Center for Pathological Diagnosis. They scanned the histopathological sections at 40x magnification with the scanner produced by Jiangfeng. For each image, there were at least two deputy chief pathologists to scrutinize it for labelling.

Using the LDCH dataset, please cite this:

[1]  Peng Jiang, Juan Liu*, Lang Wang, Jing Feng, Dehua Cao, Baochuan Pang. Classifying Cervical Histopathological Whole Slide Images via Deep Multi-Instance Transfer Learning. 2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Las Vegas, NV, USA, 2022, pp. 2302-2307. DOI: 10.1109/BIBM55620.2022.9995014. [论文下载/paper download] [数据集下载/dataset download]

The LDWBC dataset contains 22,645 images: Basophil (224), Monocyte (968), Eosinophil (539), Neutrophil (10,469) and Lymphocyte (10,445). The resolution of each image is 1280×1280 pixels. These images are collected from the samples from 150 microscope slides of 150 normal subjects. These samples are stained by Wright&Gimsa and then automatically scanned by a a light microscope with OLYMPUS BX41 and OLYMPUS Plan N 100x/1.25 to obtain corresponding digital images.

Using the LDWBC dataset, please cite this:

[1]  Hua Chen, Juan Liu*, Chunbing Hua, Zhiqun Zuo, Jing Feng, Baochuan Pang, and Di Xiao. TransMixNet: An Attenion Based Double-Branch Model for White Blood Cell Classification and Its Training with the Fuzzified Training Data. 2021 IEEE International Conference on Bioinformatics and Biomedicine (BIBM2021), pp. 842-847. 2021. Doi: 10.1109/BIBM52615.2021.9669587 [论文下载/paper download] [数据集下载/dataset download]

The LDCC dataset contains 198,952 images : Positive (60,238), Negative (25,001) and Junk (113,713), which is divided into training set and test set. The resolution of each image is 128×128 pixels. These images are collected from cervical cell samples from 2,312 participants. These samples are stained by hematoxylin and eosin (H&E) and then automatically scanned by a micro-scanning device with high resolution to obtain corresponding digital images.

Using the LDCC dataset, please cite this:

[1]  Hua Chen, Juan Liu*, Qing-Man Wen, Zhi-Qun Zuo, Jia-Sheng Liu, Jing Feng, Bao-Chuan Pang, Di Xiao. CytoBrain: Cervical Cancer Screening System Based on Deep Learning Technology[J]. Journal of Computer Science and Technology, 2021, 36(2): 347-360. Doi.org/10.1007/s11390-021-0849-3 [论文下载/paper download] [数据集下载/dataset download]

The DSCC dataset contains 15,509 images: NORMAL (5,533), ASC (3,682) and SIL (6,294). Every cell in DSCC is manually labeled by professional cytologists. And, for each cell, 78 dimensional manual features are also provided. The resolution of each image is 128×128 pixels. These images are collected from cervical cell samples from 244 participants. These samples are stained by hematoxylin and eosin (H&E) and then automatically scanned by a micro-scanning device with high resolution to obtain corresponding digital images.

Using the DSCC dataset, please cite this:

[1]  Hua Chen, Juan Liu*, Yu Jin, Baochuan Pang, Dehua Cao and Di Xiao. DSCC: A Data Set of Cervical Cell Images for Cervical Cytology Screening. International Journal of Data Mining and Bioinformatics, 2023, 27(1-3). DOI: 10.1504/ijdmb.2022.130325 [论文下载/paper download] [数据集下载/dataset download]