This is my final project for the Artificial Intelligence course at UABC. The topic to be addressed in this final project was a free criterion, that is why I chose a Deep Learning topic.
This project is about implementing Transfer Learning based on a pretrained model (MobileNet_v1_1.0_224_quantized) that classifies about 900 images; trained with ImageNet (a big Image dataset that contains about 16 million images).
I collected myself about 30 images per naturo character, and only chose 10 different characters due to a time constraint to submit the project.
I used as a template the Tensorflow Lite Android Codelab, based on that, I could modify the code and adapt it to my 10 labels to classify (images).
After I created the TensorFlowLite (.tff) model and the text file with the labels, I used the Android example app to use TensorFlowLite models, and I made a little adjustments on it.