Kotlin Backend to serve user requests for MalignantBirthmarkRecognition_Frontend.
The Backend uses Transfer Learning and the pre-trained model VGG with 16 layers from Oxford University. I provided a small shell script to make enlarge the java virtual machine boundaries to avoid OutOfMemoryExceptions.
The backend provides only a single http-post-route to upload a image. This image is then classified by the convolutional neural network and will return a json-array with all confidence-rates.
This application will be dockerized soon. Please be patient.
This project was inspired and is based on the Cat and Dog Recognizer by Klevis.
This application is open source. Please see the current license.
... are always welcome!