This project utilizes the EfficientNetB3 neural network architecture to enhance the early detection of Oral Squamous Cell Carcinoma (OSCC) through histopathological image analysis. Developed by Raja Muppidi and Bhuvana Muriki under the guidance of Professor Weihua Zhou from Michigan Tech College of Computing, this initiative aims to integrate cutting-edge AI and machine learning technologies into the field of medical diagnostics.
The focus of this project is to apply the EfficientNetB3 model to effectively classify histopathological images into two categories: OSCC and normal tissues. The model employs a sophisticated deep learning framework optimized for high accuracy and efficiency in medical image analysis.
To run this project, follow these steps:
-
Clone the repository: git clone https://github.com/your-github-username/project-repository-name.git
-
Install the required Python packages: pip install -r requirements.txt
-
Download the dataset used for training the model from Kaggle: https://www.kaggle.com/datasets/ashenafifasilkebede/dataset
Please ensure you have a Kaggle account and are logged in to access the dataset.
- Run the Jupyter notebooks in the
notebooksdirectory to train the model and evaluate its performance.
The model is based on the EfficientNetB3 architecture, utilizing layers such as:
- Convolutional and pooling layers for feature extraction.
- Dense layers with dropout regularization to prevent overfitting.
- A final softmax layer for classification.
Our model has demonstrated promising results in accurately detecting OSCC from histopathological images, achieving a high accuracy rate. Detailed results and performance metrics can be found in the results section of the Jupyter notebooks.
We extend our gratitude to Michigan Technological University for supporting this project, providing resources and an environment conducive to learning and innovation. Special thanks to Professor Weihua Zhou for his guidance and expertise.
This project is open-sourced under the MIT license. See the LICENSE file for more details.
For any queries regarding this project, please reach out to:
- Raja Muppidi (dr.rajamuppidi@gmail.com)