Improving Crowdsourcing-Based Image Classification Through Expanded Input Elicitation and Machine Learning