A set of applications (UI, Image Parser, ResNet50 model scoring API) that allow an analysis of images to obtain a "tag" for said image.

It consists of

A very basic webapp in ReactJS to upload images. This is a form to submit the file to the FileHandlerAPI and display the results of the scoring analysis.

is a golang application to get the image file and parse (standardize/decode) it to a tensor model for scoring.

a deeplearning service that uses the ResNet50 architecture to score a TF model. This is based on the work of Geert Baeke Consist of a container that uses the Azure Machine Learning computing service and a Python script that uses the Keras Applications - ResNetV2 library to receive the 4D array payload and process (classify) the image.

To see all services in action:

docker-compose up

Use the images provided in the "/example_images" folder or some of your own.