This is a project which uses deep learning algorithm to detect malaria in cell images.
The dataset used in this project is taken from here: https://lhncbc.nlm.nih.gov/LHC-research/LHC-projects/image-processing/malaria-datasheet.html
- Alexnet
- InceptionV3
- resnet101
- MobileNetV3
- Ensemble Learning Model based on InceptionV3 and MobileNetV3
Out of the all the above models, MobileNetV3 proved to be the most effective one with a training accuracy of around 94.42% and testing accuracy of around 93.40%
The deep learning model of this project is connected with a frontend webapp created with the help of NextJS via FastAPI for real time prediction. The frontend of the project is deployed on Vercel and the backend of the project is deployed on HuggingFace Spaces.
NOTE: In order to make a successful payment through Razorpay using the Card option, one can use this dummy credit card number: 4111 1111 1111 1111
- Live Preview: https://malaria-cell-classifier.vercel.app/
- Backend FastAPI link of the model: https://som11-malaria-cell-classification.hf.space/
- Swagger documentation of the FastAPI of the deep learning model: https://som11-malaria-cell-classification.hf.space/docs
- NodeJS API of the project: https://malaria-cell-detect-backend-nodejs.onrender.com/
While the model of this project can detect malaria in cell images correctly, but in some cases, the model may misclassify or fail to detect malaria altogether, therefore, it is strongly advised not to rely solely on the output of this model.