To detect diseases in chickens at an early stage using deep learning techniques, preventing mortality in chickens, farmers loss due to mortality among chickens and ultimately keeping us healthy too. The aim of this project is to develop a very intelligent system for the early identification of various diseases in chickens. VGG16 from Keras Applications was implemented for the categorical classification of "Coccidiosis" and "Healthy."
You can find the deployed link to the project here: https://chickenapp1.azurewebsites.net
(I have terminated this for now due to the high computing cost incurred in running this application.)
It is critical to follow this workflow to avoid confusion when working with each component.
-
Update config.yaml
-
Update secrets.yaml
-
Update params.yaml
-
Update the entity
-
Update the configuration manager in src config
-
Update the components
-
Update the pipeline
-
Update the main.py
-
Update the dvc.yaml
Before you run the project, make sure that you are configuring your AWS S3 Bucket called as chicken-fecal-images, that contains a zip file named chicken-fecal-images.zip that in turn has a folder named chicken-fecal-images which has two folders, namely coccidiosis and healthy. Now within each of these aforementioned folders contains the fecal images of coccidiosis infected chicken and a healthy chicken respectively.
Clone the repository
git clone https://github.com/tejangupta/Chicken-Disease-Classification.git
conda create -n chicken python=3.8 -y
conda activate chicken
pip install -r requirements.txt
export PYTHONPATH=$PYTHONPATH:/path/to/Chicken-Disease-Classification
export AWS_ACCESS_KEY_ID=<AWS_ACCESS_KEY_ID>
export AWS_SECRET_ACCESS_KEY=<AWS_SECRET_ACCESS_KEY>
python app.py
http://localhost:80/train
http://localhost:80/predict
Save the password generated under Access keys in your container registry somewhere safe. You can only see the password after enabling Admin user in Access keys.
docker build -t <login_server>/<your_project_name>:<tag> .
docker login <login_server>
docker push <login_server>/<your_project_name>:<tag>
- Build the Docker image of the Source Code
- Push the Docker image to Container Registry
- Launch the Web App Server in Azure
- Pull the Docker image from the container registry to Web App server and run