Employing mobile applications in human-machine interaction in visual pattern recognition research

Abstract

This study is part of the first author’s continued dissertation research in human-machine interaction in visual pattern recognition. Previous research focused on evaluating human-machine interaction using a flower recognition tool. Initial research showed that human interaction in color recognition improved accuracy significantly. We then looked more deeply into various automated color recognition algorithms and ways of combining them with human feedback. Described here is the process of upgrading the initial system into a new mobile application using Appinventor. After data collection, models were built for various color spaces. Sharing this experience may help other researchers incorporating a human-computer interaction component into their work.