vCanteen: A Smart Campus Solution to Elevate University Canteen Experience
B. Vatcharakomonphan, Chompakorn Chaksangchaichot, N. Ketchaikosol, T. Tetiranont, T. Chullapram, P. Kosittanakiat, P. Masana, P. Chansajcha, S. Suttawuttiwong, S. Thamkittikhun, S. Wattanachindaporn, A. Boonsith, Chotirat Ann Ratanamahatana, N. Prompoon, Manisa Pipattanasomporn · 2019
The persistent problem circulating around various university canteens has always been about high crowd density during lunch hours. To efficiently tackle this issue, a platform called “vCanteen” has been developed that integrates an online food ordering system, a virtual queuing system, together with a machine learning-based crowd estimation system. vCanteen aims at reducing queuing time when ordering food, and allowing users to know the estimated crowd density in a university canteen in real-time. The crowd estimation system has been developed using a multi-column convolutional neural network (MCNN). This paper discusses the vCanteen prototype that was developed and tested at the canteen in the Faculty of Engineering, Chulalongkorn University, Thailand. The description of the crowd estimation system is provided in details including error evaluation and lessons learned.