Irregular luggage classification experiments using data from Tromsø airport

Abstract

This paper examines the current state-of-the-art in object tracking and detection in the context of baggage handling systems using a variety of technologies, with special interest in image classification. Data collection at Tromsø Airport using simple and inexpensive equipment in order to capture images of different types of luggage will be described. Furthermore, a selection of three relevant techniques for image classification will be investigated, and experiments will be conducted to detect irregular and potentially problematic bags in the collected dataset. Finally, recommendations for use cases and an interface to an intelligent baggage handling system will be proposed.