IMPLEMENTASI ARSITEKTUR REST DALAM APLIKASI VIDEO CONFERENCE DENGAN FITUR PENGENALAN EMOSI MENGGUNAKAN WEBRTC

• Main Author: Muhammad Reynaldi, - (Author)
• Format: Book
• Published: 2023-08-15.
• Subjects: Thesis; NonPeerReviewed
• Online Access: Link Metadata
• Notes: http://repository.upi.edu/100004/1/S_RPL_1904819_Title.pdf; http://repository.upi.edu/100004/2/S_RPL_1904819_Chapter1.pdf; http://repository.upi.edu/100004/3/S_RPL_1904819_Chapter2.pdf; http://repository.upi.edu/100004/4/S_RPL_1904819_Chapter3.pdf; http://repository.upi.edu/100004/5/S_RPL_1904819_Chapter4.pdf; http://repository.upi.edu/100004/6/S_RPL_1904819_Chapter5.pdf; http://repository.upi.edu/100004/7/S_RPL_1904819_Appendix.pdf
• Keywords: QA76 Computer software; T Technology (General)

Emosi memiliki pengaruh yang signifikan dalam proses pembelajaran, karena dapat memengaruhi ingatan dan tindakan. Saat ini, berbagai jenis Application Programming Interface (API) pengenalan emosi telah tersedia, memberikan kesempatan untuk menerapkan aplikasi pengenalan emosi dan aplikasi visualisasi secara real-time. Mendapatkan data real-time biasanya dilakukan melalui penggunaan suatu API, seperti Representational State Transfer (REST). Terdapat beberapa aplikasi yang digunakan untuk mengenali emosi siswa dalam pembelajaran daring, salah satunya adalah aplikasi video conference. Salah satu teknologi yang digunakan dalam membangun suatu aplikasi video conference yaitu WebRTC. Penelitian ini bertujuan untuk mengembangkan aplikasi WebRTC pengenalan emosi dengan arsitektur REST, serta menganalisis performa aplikasi back-end dan front-end. Performa aplikasi back-end diukur menggunakan metrik Quality of Service (QoS) seperti response time, throughput, memory utilization, dan CPU Load. Performa aplikasi front-end diukur dengan metrik Google Lighthouse Performance. Hasilnya pada aplikasi back-end pada endpoint Recognition Grup nilai rata rata Response Time sebesar 2567,34 ms, Throughput 36,89 request/s, Memory Utilization 622,05 MB, CPU Load 9,53 %. Sedangkan pada endpoint Recognition Individu nilai rata rata Response Time sebesar 3209,18 ms, Throughput 29,39 request/s, Memory Utilization 623,96 MB, CPU Load 7,67 %. Hasilnya pada aplikasi front-end nilai rata - rata pada metrik FCP sebesar 936,1 ms, SI sebesar 1095,28 ms, LCP sebesar 1154,54 ms, TTI sebesar 972,55ms, TBT sebesar 0,728 dan CLS sebesar 0. Dengan demikian keseluruhan metrik menghasilkan nilai 96% Performance Score sehingga performa aplikasi front-end dapat dikatakan Baik. -------- Emotions have a significant impact on the learning process as they can influence memory and actions. Currently, various types of Emotion Recognition Application Programming Interfaces (APIs) are available, providing opportunities to implement real-time emotion recognition and visualization applications. Real-time data acquisition is typically accomplished through the use of an API, such as Representational State Transfer (REST). One of the applications used for recognizing students' emotions in online learning is video conferencing applications. WebRTC is one of the technologies used to build a video conferencing application. This research aims to develop a WebRTC-based emotion recognition application with a REST architecture and analyze the performance of the backend and front-end components. The back-end application's performance is measured using Quality of Service (QoS) metrics, including response time, throughput, memory utilization, and CPU load. The front-end application's performance is measured using Google Lighthouse Performance metrics. The results for the back-end application show that, in the Recognition Group endpoint, the average response time is 2567.34 ms, throughput is 36.89 requests/s, memory utilization is 622.05 MB, and CPU load is 9.53%. Meanwhile, in the Recognition Individual endpoint, the average response time is 3209.18 ms, throughput is 29.39 requests/s, memory utilization is 623.96 MB, and CPU load is 7.67%. The results for the front-end application indicate that the average values for the FCP, SI, LCP, TTI, TBT, and CLS metrics are 936.1 ms, 1095.28 ms, 1154.54 ms, 972.55 ms, 0.728, and 0, respectively. Thus, the overall metrics yield a 96% Performance Score, indicating that the front-end application's performance can be considered Good.