I created this application for aquarists who need to calculate the optimal lighting for their tanks.
In the practical part the software capable of determining variety of apples with web camera in real time using convolution neuron network was developed. Created dataset contains 5200 images in ratio 80/20 for training and validation set. After training of neuron network we obtained almost 100% of apples variety recognition.
This application developed in apache cordova and builded only for android OS. I wrote data to JSON files. The App developed for colleagues (department of planners) in previous employment. Try install on your mobile :).