Multimedia › AiCamera
thumb_up Follow me on GitHub to get notifications about my projects (like Fossdroid Core), thanks! close


Camera that simulates depth of field using a neural network
Version: 1.0
Added: 02-07-2019
Updated: 02-07-2019
This is a camera app for Android that automatically recognises the pixel location of people, cats or dogs in the camera view and blurs the background simulating depth of field. It can also gray out the background leaving only the recognised object in colour.
The app uses a convolutional neural network to perform the detection. A TensorFlow implementation of the neural network can be found here. It was trained on a dataset of images consisting of people, cats and dogs and the trained model then exported as a protobuf file which contains the graph definition and the trained weights.
Screenshot of AiCamera Screenshot of AiCamera
code Source file_download Download