Personalized object recognition for augmenting human memory
Hosub Lee, Cameron Upright, Steven Eliuk, Alfred Kobsa · 2016
We propose a novel wearable system that enables users to create their own object recognition system with minimal effort and utilize it to augment their memory. A client running on Google Glass collects images of objects a user is interested in, and sends them to the server with a request for a machine learning task: training or classification. The server processes the request and returns the result to Google Glass. During training, the server not only aims to build machine learning models with user generated image data, but also to update the models whenever new data is added by the user. Preliminary experimental results show that our system DeepEye is able to train the custom machine learning models in an efficient manner and to classify an image into one of 10 different user-defined categories with 97% accuracy. We also describe challenges and opportunities for the proposed system as an external memory extension aid for end users.