The project consists of automated image based malaria detection system using a custom Convolutional Neural Network (CNN). The model classifies microscopic blood smear images into two categories: Parasitized and Uninfected. The objective is to support faster and more reliable diagnosis in clinical and resource-limited settings by utilizing deep learning. Manual examination of blood slides is very time consuming operation and required the expertise of skilled professionals in the respective domains. By using CNN model this project evaluates visual data, identifying siginificant patterns, and forecast infection status with high precision. It also helped as digital AI based diagonistic and hence make it more relevant in healthcare practises and minimizing human errors.
- Automated Image Classification
- Efficient Preprocessing
- Custom CNN Architecture
- Performance Montioring
- Model Training
To run this project, you'll need the following software and libraries:
- Python 3.8+
- Jupyter Notebook
- Pandas
- NumPy
- Scikit-learn
- Statsmodels
- TensorFlow
- Matplotlib
- Seaborn
Data Preprocessing: Run the preprocessing to clean the data.
Exploratory Data Analysis (EDA): Use the notebooks in the notebooks/ directory to conduct EDA and visualize feature relationships.
Model Implementation and Evaluation: Implement and evaluate models using the provided notebooks and scripts.
Feature Engineering: Investigate the impact of various features and create new ones using the feature engineering scripts.
Model Tuning and Validation: Tune hyperparameters and validate models to ensure optimal performance.