Digital transformation in higher education is driving the uptake of online tests, which require academic integrity, security, and robust user experience. In the context of authentication of users, deeplearningbased face recognition, in particular the Convolutional Neural Network (CNN) architectures, such as MobileNetV2, combined with intermediate filter, promises to deliver a consistent performance across a wide range of devices and imaging environments. However, there are limited comprehensive studies evaluating the final integration of the median filter and MobileNetV2 in high-value test scenarios. This study contributes by proposing an effective end-to-end Face Authentication Pipeline, assessing the median impact of filtering on MobileNetV2 performance, and validating it with a prototype application. The authentic face dataset was collected using the Teachable Machine, preprocessed with cropping, resizing, and median filtering, and then augmented through rotation, shift, shear, zoom, reversal, and brightness adjustment. The MobileNetV2 model was trained with Adam in a stepwise manner, starting with 0.001 and then 0.0001 for 20 epochs in a batch size of 32, and was evaluated for accuracy, precision, recall, and F1 score. Results show that the accuracy curve has remained stable at almost 95 percent during the 20th epoch; most grades achieved 1.00 in both precis, recall and F1, with some classings showing a limited decrease due to facial similarity or expression differences. These findings confirm that MobileNetV2 median filtering canbe the basis for an effective, accurate and ready to integrate face recognition in online testing applications on a wide range of devices

Perkembangan teknologi kecerdasan buatan (AI) dan visi komputer telah membuka peluang besar dalam penerapan pengenalan ekspresi wajah pada berbagai bidang. Dalam konteks pendidikan tinggi, keterlibatan mahasiswa selama proses belajar menjadi faktor penting yang masih sulit diukur secara objektif menggunakan metode konvensional. Namun pada kenyataannya, penelitian sebelumnya masih jarang menguji performa arsitektur CNN populer secara langsung di lingkungan pembelajaran nyata dengan kondisi pencahayaan dan pose yang beragam. Penelitian ini berkontribusi dengan membandingkan kinerja dua arsitektur deep learning, yaitu VGG-16 dan ResNet, dalam klasifikasi ekspresi wajah mahasiswa pada Smart Learning Environment. Penelitian dilakukan dengan pendekatan eksperimen kuantitatif melalui lima tahapan, yaitu pengumpulan data wajah mahasiswa di kelas, preprocessing berupa cropping, resizing, dan augmentasi, pengembangan model CNN, pelatihan menggunakan data split 80% training dan 20% validasi, serta evaluasi dengan metrik akurasi, presisi, recall, dan F1-score. Hasil eksperimen menunjukkan bahwa VGG-16 unggul dalam mengenali ekspresi suka dengan nilai F1-score tertinggi sebesar 85%, sedangkan ResNet relatif lebih baik pada ekspresi bosan dengan F1-score 73,2%. Sementara itu, keduanya sama-sama lemah dalam mengenali ekspresi tidak suka. Temuan ini mengimplikasikan bahwa VGG-16 lebih sesuai digunakan untuk mendukung analisis keterlibatan mahasiswa secara real-time dalam Smart Learning Environment berbasis AI.