<![CDATA[Abstrak - Dalam era teknologi yang semakin maju, kebutuhan akan sistem absensi yang efisien dan otomatis menjadi sangat penting, terutama di lingkungan kerja dengan aktivitas padat seperti Toko Roti Bread Smile. Absensi manual yang masih digunakan terbukti menimbulkan berbagai kendala, seperti keterlambatan rekapitulasi data, risiko kehilangan dokumen, serta kesulitan pemantauan kehadiran secara real-time. Untuk mengatasi permasalahan tersebut, dikembangkan sistem absensi berbasis Internet of Things (IoT) dengan penerapan teknologi Face Recognition dan pendekatan Edge Computing. Sistem ini menggunakan dengan arsitektur MobileNetV2 yang ringan dan akurat dalam mengenali wajah secara real-time. Proses inferensi dilakukan langsung pada perangkat ESP32 tanpa ketergantungan pada koneksi internet, sehingga tetap dapat digunakan dalam kondisi offline. Sistem juga memanfaatkan kamera OV2640 untuk menangkap citra wajah dan layar TFT sebagai antarmuka yang menampilkan hasil pengenalan secara langsung kepada pengguna. Selain itu, sistem dilengkapi dengan antarmuka web yang memungkinkan admin untuk memantau, mengelola, dan merekap data absensi secara efisien dan terpusat. Sistem dirancang agar dapat mendeteksi wajah dengan cepat dan akurat dalam berbagai kondisi pencahayaan dan orientasi wajah. Evaluasi dilakukan melalui pengujian pengenalan wajah serta sinkronisasi data antara mode offline dan online. Hasil menunjukkan bahwa sistem mampu mencatat kehadiran secara tepat, aman, dan dapat diandalkan. Pendekatan ini terbukti efektif dalam mendukung pengelolaan absensi yang modern, fleksibel, serta sesuai diterapkan pada lingkungan kerja berskala kecil hingga menengah.Kata kunci : Absensi, Internet of Things; Face Recognition; Convolutional Neural Network; Edge Computing; ESP32; Abstract - In an era of increasingly advanced technology, the need for an efficient and automated attendance system is very important, especially in a work environment with busy activities such as Bread Smile Bakery. The manual attendance that is still used has proven to cause various obstacles, such as delays in data recapitulation, the risk of losing documents, and difficulties in monitoring attendance in real-time. To overcome these problems, an Internet of Things (IoT)-based attendance system was developed with the application of Face Recognition technology and the Edge Computing approach. This system uses Convolutional Neural Network (CNN) algorithm with MobileNetV2 architecture that is lightweight and accurate in recognizing faces in real-time. The inference process is carried out directly on the ESP32 device without dependence on an internet connection, so it can still be used in offline conditions. The system also utilizes an OV2640 camera to capture facial images and a TFT screen as an interface that displays the recognition results directly to the user. In addition, the system is equipped with a web interface that allows admins to monitor, manage, and recap attendance data efficiently and centrally. The system is designed to detect faces quickly and accurately in various lighting conditions and face orientations. Evaluation is conducted through face recognition testing as well as data synchronization between offline and online modes. The results show that the system is able to record attendance precisely, securely, and reliably. This approach is proven to be effective in supporting modern, flexible attendance management, and is suitable for small to medium-sized work environments. Keywords: Attendance; Internet of Things; Face Recognition; Convolutional Neural Network; Edge Computing; ESP32;]]>
