Related Experiment Video
Updated: Aug 5, 2026
05:33
Introduction of an Integrated Pathology Image Management, Artificial Intelligence, and Reporting System
Published on: July 11, 2025
BiCervi: Pap smear microscopy image dataset for cervical cancer detection and classification
Hope Mbelwa1, Judith Leo1, Crispin Kahesa2
1School of Computational and Communication Science and Engineering, The Nelson Mandela African Institution of Science and Technology, Arusha, 23311, Tanzania.
Data in Brief
|July 30, 2026
Summary
This study introduces a new dataset of 3,000 Pap smear images from Tanzania to aid cervical cancer screening. This resource supports the development of AI tools for early detection, especially in low-resource settings.
Area of Science:
- Medical imaging
- Computational pathology
- Public health
Background:
- Cervical cancer remains a significant global health issue, particularly in low-resource regions.
- Manual Pap smear analysis faces challenges with specialist availability and inter-observer variability.
- Limited availability of African Pap smear datasets hinders machine learning model development.
Purpose of the Study:
- To present a novel, curated dataset of Pap smear microscopy images from Tanzania.
- To facilitate the development and evaluation of AI-driven cervical cancer detection tools.
- To address the scarcity of relevant data for machine learning in African screening contexts.
Main Methods:
- Collected 3,000 Pap smear images using a digital microscope (10x/0.25 objective) at The Ocean Road Cancer Institute, Tanzania.
- Images stored as RGB JPEG at 2592 x 1944 resolution.
- Utilized The Bethesda System for labeling into eight diagnostic categories, with independent expert review and adjudication for quality control.
Main Results:
- A dataset of 3,000 labeled Pap smear images is now available.
- Images cover eight diagnostic categories, including NILM, ASC-US, ASC-H, AGC, LSIL, HSIL, Squamous Cell Carcinoma, and Adenocarcinoma.
- The dataset underwent rigorous quality control, ensuring label consistency and reliability.
Conclusions:
- The curated Tanzanian Pap smear dataset is a valuable resource for advancing AI in cervical cancer diagnostics.
- This dataset can accelerate the creation of AI-based decision support systems for cytology screening.
- Enables more equitable development and validation of machine learning models for cervical cancer detection globally.
