Jaringan Syaraf Tiruan dan Face Landmark untuk Deteksi Bicara Antar Peserta Ujian Melalui Gerakan Mulut

Putra, Andra Kurnia (2023) Jaringan Syaraf Tiruan dan Face Landmark untuk Deteksi Bicara Antar Peserta Ujian Melalui Gerakan Mulut. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 02311940000087-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (2MB) | Request a copy

Abstract

Pembelajaran di perkuliahan telah mengalami transisi melalui aplikasi mobile yang saat ini mulai banyak digunakan oleh universitas untuk memberikan pelayanan dan aktivitas pembelajaran. Transisi ini disebut dengan istilah mobile learning. Penggunaan berbagai macam media mobile learning tidak hanya sebagai media memberi dan menerima ilmu, tetapi juga dapat dimanfaatkan sebagai sarana melaksanakan ujian secara daring. Ujian yang dilakukan secara daring membuka celah bagi peserta ujian untuk melakukan kecurangan karena tidak ada pengawas yang begitu ketat seperti ujian pada umumnya. Kecurangan yang difokuskan pada penelitian ini adalah bicara antar peserta ujian. Berbagai macam penyedia layanan ujian daring tersedia, akan tetapi layanan ini membutuhkan perangkat khusus yang memadai dan membutuhkan manusia untuk mengoperasikannya. Penelitian ini berfokus pada pengawasan deteksi membuka mulut. Mulut terbuka ini dapat dideteksi dengan melihat perbandingan antara panjang mulut secara vertikal terhadap panjang mulut secara horizontal. Penelitian ini menggunakan metode Convolutional Neural Network (CNN) dan facial landmark untuk deteksi bicara tersebut. Pengambilan data dilakukan dengan perekaman selama ujian melalui aplikasi zoom dengan jumlah sampel sebanyak 10. Ujian yang dilakukan berupa penampilan 10 soal. Diperoleh hasil akurasi sebesar 98% dari deteksi yang diperoleh dari perhitungan kesesuaian antara skenario bicara yang diatur terhadap hasil deteksi program.
=================================================================================================================================
Learning in lectures has undergone a transition through mobile applications which are now widely used by universities to provide services and learning activities. This transition is called mobile learning. The use of various kinds of mobile learning media is not only a medium of giving and receiving knowledge, but can also be used as a means of carrying out online exams. Exams conducted online open up loopholes for examinees to cheat because there is no proctor who is as strict as exams in general. The cheating that this study focused on was communication between examinees. A wide variety of online exam service providers are available, but these services require adequate specialized devices and require humans to operate. This study focused on oral opening detection surveillance. This mouth opening can be detected by looking at the ratio between the length of the mouth vertically to the length of the mouth horizontally. This study used Convolutional Neural Network (CNN) methods and landmark facials to speech detection. Data collection was carried out by recording during the exam via the zoom application with a total of 10 samples. The exam conducted is in the form of a 10-question performance. Accuracy results of 98% of the readings obtained from the calculation of conformity between regulated communication scenarios and program readings
were obtained

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Mobile Learning, Convolutional Neural Network, Facial Landmark
Subjects:	L Education > LB Theory and practice of education > LB1044.87 Internet in education (e-learning). Virtual reality in education. Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User:	Andra Kurnia Putra
Date Deposited:	11 Jan 2024 07:03
Last Modified:	11 Jan 2024 07:03
URI:	http://repository.its.ac.id/id/eprint/103592

Actions (login required)

View Item