Pneumonia Detection using Local Search aided Sine-Cosine Algorithm

Based on our paper "Pneumonia Detection from Lung X-ray Images using Local Search Aided Sine Cosine Algorithm based Deep Feature Selection Method" published in International Journal of Intelligent Systems, Wiley (2021).

Note: Access the preprint here. For the PDF of the published version of the manuscript, please email the first and/or second authors at: [email protected] and [email protected].

Model Architecture

Requirements

To install the required dependencies run the following in command prompt: pip install -r requirements.txt

Running the codes:

Required directory structure:

(Note: train and val contains subfolders representing classes in the dataset.)

+-- data
|   +-- .
|   +-- train
|   +-- val
+-- AbSCA.py
+-- local_search.py
+-- main.py

Then, run the code using the command prompt as follows:

python main.py --data_directory "data"

Available arguments:

• --epochs: Number of epochs of training. Default = 20
• --learning_rate: Learning Rate. Default = 0.001
• --batch_size: Batch Size. Default = 4
• --momentum: Momentum. Default = 0.9

Citation:

If this article helps in your research in any way, please cite us using:

@article{chattopadhyay2022pneumonia,
  title={Pneumonia detection from lung X-ray images using local search aided sine cosine algorithm based deep feature selection method},
  author={Chattopadhyay, Soumitri and Kundu, Rohit and Singh, Pawan Kumar and Mirjalili, Seyedali and Sarkar, Ram},
  journal={International Journal of Intelligent Systems},
  DOI={10.1002/int.22703},
  volume={37},
  number={7},
  pages={3777--3814},
  year={2022},
  publisher={Wiley Online Library}
}