Santa detector

The instructions for this project are described in the following steps:

1. Build a deep learning model in Keras to classify images as Santa or not-Santa. A simple CNN model is used. Transfer learning with the VGG16 model is also shown. https://github.com/salekd/santa/wiki/Image-classification-with-Keras
2. Deploy a serverless function with OpenFaaS, using the Keras classifier and starting from the ubuntu:16.04 Docker image. Tips on how to integrate this function with a security camera are also given. https://github.com/salekd/santa/wiki/OpenFaaS

I also tried building a more sophisticated classifier using TensorFlow Object Detection API and transfer learning with a pre-trained SSD MobileNet v1 model. Although I managed to train, it does not perform well due to an insufficient dataset. The whole endeavour is described here https://github.com/salekd/santa/wiki/Transfer-learning-with-SSD-MobileNet-v1