A smart identification system based on computer vision for supporting hotel check-in process

Citizen ID card verification is critical across various sectors, including banking, access control, product pickups, and hotel check-ins. However, the traditional manual registration process is time-consuming and error-prone. Moreover, the prevalence of mask usage during the COVID-19 pandemic presents challenges for face recognitionsystems, leading to potential inconveniences and delays. To address these concerns, this study introduces two advanced models: Optical Character Recognition (OCR) and face-mask recognition. The OCR model accurately extracts information from ID cards and automatically populates the registration form, while the face-mask recognition model improves face-recognition accuracy, even with individuals wearing masks.By implementing these models, check-in procedures can be streamlined, aiming toenhance efficiency, data accuracy, and potentially improve customer satisfaction.

公民身份证核验在银行业、门禁管理、物品领取以及酒店入住等诸多领域均至关重要。然而传统的手动登记流程不仅耗时耗力，还极易出现差错。此外，新冠疫情期间口罩佩戴的普及给人脸识别系统带来了新的挑战，进而可能引发流程不便与延误。为解决上述问题，本研究引入了两种先进模型：光学字符识别（Optical Character Recognition, OCR）与口罩人脸识别模型。其中，OCR模型可精准提取身份证内的证件信息，并自动完成登记表单的填写；口罩人脸识别模型则能在人员佩戴口罩的场景下，提升人脸识别的准确率。通过部署上述模型，可优化各类核验登记流程，旨在提升业务办理效率与数据准确性，同时有望进一步提升用户满意度。