Toward Affordable and Practical Home Context Recognition: -Framework and Implementation with Image-based Cognitive API-

Abstract

To provide affordable context recognition for general households, this paper presents a novel technique that integrate image-based cognitive Application Program Interface (API) and light-weight machine learning. Our key idea is to regard every image as a document by exploiting “tags” derived by the API. We first present a framework that specifies a common workflow of the machine-learning-based home context recognition. We then propose a pragmatic method that implements the framework using the “image-as-a-document” approach.