Browse Datasets

Nigerian Colposcopy Images

Cervical Cancer Screening Images from Nigeria

image-classification computer-aided-diagnosis computer-vision 3,356 images CC BY-NC 4.0
Access Dataset Read Paper

Description

Multi-magnification digital colposcopy images from women at a Nigerian hospital, captured before and after acetic-acid application, for computer-aided cervical cancer screening. Data files are hosted on Mendeley Data (doi:10.17632/j5mhfhxrwj.1).

Quick Start

Python 
# Download the images from Mendeley Data (CC BY-NC 4.0):
#   https://data.mendeley.com/datasets/j5mhfhxrwj/1
from torchvision.io import read_image

Project Ideas

Lesion Classifier

Train a lightweight CNN to flag normal vs abnormal cervical lesions, optimising inference for low-resource clinics.

Intermediate

Pre/Post Acetic Acid

Compare classification accuracy on pre- vs post-acetic-acid images.

Intermediate

Ethical Considerations

  • Highly sensitive medical imagery requiring strict privacy handling.
  • False negatives in oncology carry severe consequences; communicate model limits clearly.