This project aims to detect anemia based on images of the conjunctiva (the inner part of the lower eyelid). Anemia is a condition characterized by a lack of healthy red blood cells, and it can often be detected by examining the color of the conjunctiva. Using a Convolutional Neural Network (CNN) implemented in TensorFlow, this model classifies images into anemic and non-anemic categories.
The dataset used for this project is sourced from Kaggle and can be found here. The dataset consists of augmented images of the conjunctiva, labeled as anemic or non-anemic.
The model is built using a Convolutional Neural Network (CNN) architecture, which includes:
- Convolutional layers for feature extraction
- Max-pooling layers for down-sampling
- Fully connected layers for classification
Data augmentation techniques such as rotation, zoom, and flip were applied to enhance the training process and prevent overfitting.
The model was trained using TensorFlow. Key details include:
- Loss function: Binary Crossentropy
- Optimizer: Adam
- Metrics: Accuracy
- Training epochs: 100
- Validation split: 20%
The model achieved an accuracy of 91% on the validation set. Below are some sample predictions and the accuracy plot over the epochs.
The plot shows the training and validation accuracy over 100 epochs, demonstrating the model's learning progression.
To use this model, clone the repository and ensure you have the necessary dependencies installed:
git clone https://github.com/aymen-000/AIQuest_DATATHON.git
cd AIQuest_DATATHON
pip install -r requirements.txt