You are to write a GUI Java/Kotlin application that run on the Android platform and can run under an Android Emmulator using Android Java IDE and Tensorflow API that features some Computer Vision Application. Your Application MUST BE APPROVED by the instructor and should be of significant complication to be interesting and have some barrier of entrance (meaning would take at the least a computer vision scientist to implement it). You are going to propose a Mobile Imaging Application that in some way will ASSIST One of the following groups:

• a blind or low-vision (not completely blind) person
• senior citizen
• someone with reduce physical mobility
• an animal
• children
• athletes
• illness situation awareness communities

 

YOU will be assigned to a group that will be in one of these designated areas --- if you choose an animal you must have access to this kind of animal for testing (e.g. if this is for dogs, you must have a dog). It has to be a real project solving some kind of real problem. It is considered Assistive Technology

 

this will involve self learning of Android Programming ---you have my Mobile Programming website to help you with general concepts. BUT MOST OF THE MATERIAL YOU WILL BE LEARNING from the Android developer website directly o your own. and we will have some discussions in class.

 

 

IMPORTANT REQUIREMENT:

You MUST use in your application the Google TensorFlow API (and also read here and here) and use some kind of Deep Learning Network (CNN, LSTM, etc) to implement your recognition layer and this requires Asynchronous calling.Note: for this project you are NOT allowed to use a pre-existing trained Deep Learning Network for your main network (or course you can retrain one with your dataset and your classes suited to your problem )

*********************************************************************

you can do both or only one or none of the following extra credit options

EXTRA CREDIT A- you will get 30 extra points if you use Google TensorFlow and train more than one model (must make sense in your system...discuss with instructor).
see this video for mobile and backend tensor flow

EXTRA CREDIT B- you will get 10 extra points if you some how additionally make use Google Cloud Vision or Google Cloud Machine Learning API but, not to replace the main Deep Learning Network you are required to use in your application. (which means you will make an Asynchronous call passing your data to Google Cloud Vision service --you must register)

EXTRA CREDIT C- you will get 30 extra points if additional Hardware sensor like a IR camera attached to device (see instructor for approval)

EXTRA CREDIT D- you will get 15 extra points if you train on the Cloud (not colab) AND create documents on how to train on cloud and share and discuss with class PRIOR to due date of project --so that others can replicate this for their projects