This repository contains the implementation of deep learning Convolutional Neural Network (CNN) algorithms for cervical cancer screening. The algorithm aims to assist in the early detection and classification of cervical cancer from digital cervical images. We trained one from scratch, and then we also tried transfer learning by training on top of VGG16 and ResNet50, but this never improved the performance.
Cervical cancer is a significant global health issue, particularly in low-resource settings. Early detection and accurate classification of cervical abnormalities are crucial for effective treatment and improved patient outcomes. This repository presents a deep learning algorithm that utilizes a CNN to analyze digital cervical images and identify potential cancerous or pre-cancerous regions.
Introduction
System Requirements
Required Dependencies
Dataset
The algorithm was developed in Python, utilizing;
- TensorFlow 2.11.0
- Keras 2.12.0
- MSI GL75 Leopard 10SFR laptop
- 8GB NVIDIA RTX 2070 GDDR6 Graphical Processing Unit (GPU)
- CUDA 12.1
- cuDNN SDK 8.7.0
The CUDA 12.1 and cuDNN 8.7.0 can be downloaded from the official NVIDIA Website.
NumPy
Pandas
Tensorflow
OpenCV
The dataset used for training and evaluation can be obtained from Kaggle. It consists of a collection of digital cervical images labelled with corresponding class annotations.