Dogs were found to be the most popular pets in India, comprising 68 percent of pets owned, according to a January 2022 survey conducted by Rakuten Insight. Research has repeatedly demonstrated the significant role that pets play in an individual's life. Studies have indicated that pet ownership can lead to increased physical activity, lower serum cholesterol and triglyceride levels, and a reduced likelihood of cardiovascular events. Furthermore, owning a pet has been linked to lower levels of depression and mental stress, and greater self-esteem. While dogs can provide emotional and mental health benefits, they are also susceptible to various diseases and infections, which can pose a risk to certain groups such as young children, pregnant women, the elderly, and immunocompromised individuals. Direct or indirect contact with dogs has been identified as a risk factor for numerous zoonotic infections caused by bacteria, fungi, parasites, and viruses.
This project aims to predict various types of dog skin diseases, a zoonotic disease, by utilizing diverse machine-learning models and evaluating their performance based on various parameters. The ultimate goal is to detect the type of disease and offer recommendations for precautions that should be taken to prevent the spread of the disease.
The project consists of the following components:
- A dataset of dog skin disease images
- Jupyter notebooks for training and evaluating models
- Python scripts for preprocessing data, training models, and generating predictions
- Django apps to create webapp for the project
- Dockerfile for building a Docker container
- Readme file for project documentation
- python==3.9
- Django
- tensorflow
- keras==2.11.0
- opencv-python==4.7.0.72
- numpy==1.21.5
- django-compressor
The dataset consists of images of various dog skin diseases available on internet. The dataset is split into training, validation, and testing sets. The training set will be used to train the model, the validation set will be used to tune hyperparameters, and the test set will be used to evaluate the final model.
The images will be resized to a standard size of 224x224 pixels and normalized to have pixel values between 0 and 1. The data will also be split into training, validation, and testing sets.
The deep learning models is built using TensorFlow and Keras libraries. The following models will be trained and evaluated:
- InceptionV3
- MobileNetV2
The models will be trained on the training data, and their performance will be evaluated on the validation data. The best performing model will be selected for deployment.
The selected model will be deployed using a Docker container. The Dockerfile will include all the necessary dependencies to run the model, including the TensorFlow and Keras libraries.
Docker image: vrukshali26/wecareforfurs:v2
Run the container:
docker run -dit -p 8000:8000 --name=<name_of_container> -v <location_store_images>:/home/app/webapp/media/ vrukshali26/wecareforfurs:v2
Fur-video.mp4
This project aims to develop a deep learning model that can accurately detect and classify dog skin diseases based on images. The model has trained on a publicly available dataset and achieved high accuracy on the testing set. The trained model can be used in veterinary clinics to assist veterinarians in the early detection and diagnosis of skin diseases in dogs.
Overall, this project demonstrates the potential of deep learning in improving animal healthcare and emphasizes the importance of utilizing advanced technology to support veterinary professionals in their daily work.
This project is licensed under the MIT license. Feel free to use it. Thanks!