A Survey on Attendance Automation, Menu-Based Nutrition Tracking, and Predictive Planning in Institutional Mess Systems: Gaps and Opportunities | IJCSE Volume 9 – Issue 6 | IJCSE-V9I6P13

The operational efficiency of institutional food services is significantly impaired by unreliable diner headcount and meal planning models that use static, historical data, which leads to substantial food waste. Although the literature indicates the maturity of high-fidelity Face Recognition (FR) for attendance [1]-[3], [5], [6], [7], [8] and the ability of Machine Learning (ML) models to forecast demand, these fields remain functionally siloed. This paper identifies the important research gap of investigating a unified, low-latency data pipeline to translate real-time high resolution, biometric attendance data into direct operational planning input. We organize the component technologies into three layers – Perception, Logistics, and Analytics, – proposing a Reference Architecture for a Smart Mess System (SMS). The proposed architecture articulates the necessary functional and data linkages for effectively automating nutrition logging, mitigation of waste via predictive planning, and true efficiency in resource allocation in institutional food service.

This study confirms there is a clear research gap in the development of unified Smart Mess Systems (SMS) that use high accuracy Face Recognition (FR) specifically contextualized to mess environments. The proposed SMS Reference Architecture gives a solid conceptual scaffolding to harness the existing maturity of computer vision for perception and scalable data management tools for prediction. The SMS establishes a reliable data pipeline that connects biometric attendance to robust predictive models, providing an innovative, operationally superior offering to mitigate food waste and improve food service management in higher volume institutional settings. Future areas of research should prioritize: FR robustness in untethered settings: to make the models robust against complex real-world variables such as interference of lighting conditions, flows of many students, and different levels of occlusion in the mess. Closed-loop validation: use sensor based post-consumption waste monitoring for empirical data collection [4] to verify the validity and update of predictive ML models, thereby creating a true closed-loop system for optimization.

[1]M. Gopila and D. Prasad, “Machine learning classifier model for attendance management system,” in Proc. Fourth Int. Conf. 1-SMAC (IoT in Social, Mobile, Analytics and Cloud) (1-SMAC), Palladam, India, 2020, pp. 280-285. DOI: 10.1109/I-SMAC49090.2020.9243363. [2]R. A. A. Helmi, S. S. E. Yusuf, M. I. B. Abdullah, and A. Jamal, “Face Recognition Automatic Class Attendance System (FRACAS),” in Proc. Int. Conf. Autom. Control Intell. Syst. (ICACIS), Selangor, Malaysia, 2019, pp. 1-6. DOI: 10.1109/ICACIS49397.2019.9015096. [3]S. Matilda and K. Shahin, “Student Attendance Monitoring System Using Image Processing,” in Proc. Int. Conf. Syst. Comput. Autom. Netw., 2019, pp. 1-6. DOI: 10.1109/ICSCAN.2019.8878806. [4]A. Utsav, A. Kumar, A. Kumari, S. Awasthi, and M. A. Rahman, “IoT Enabled Smart Mess with Nutrition Benefits and Waste Management,” in Proc. 4th Int. Conf. Adv. Electron. Commun. Eng. (AECE), 2024, pp. 1-6. DOI: 10.1109/AECE59216.2024.10543204. [5]A. Rao, “AttenFace: A Real Time Attendance System using Face Recognition,” arXiv preprint arXiv:2211.07582, 2022. [6]J. B. Jai and J. C. U. et al., “Intelligent Face Tracking Attendance System Using LBPH and Kalman Filtering,” Granthaalayah Publication, 2024. DOI: 10.29121/granthaalayah.v12.i6.2024.5822. [7]T. Goswami et al., “Attendance Monitoring System using Facial Recognition,” IJERT, 2020. (Note: No standard DOI is assigned; the Paper ID is IJERTCONV8IS10059). [8]Thirukrishna et al., “Smart Attendance System Using Face Recognition,” AJEAT, 2023. DOI: 10.51983/ajeat-2023.12.2.3968.