Ananda Myzza Marhelio, - and Munir, - and Yaya Wihardi, - (2025) DETEKSI FOKUS ATENSI VISUAL PESERTA DIDIK DI RUANG KELAS BERDASARKAN POSE KEPALA MENGGUNAKAN EFFICIENTNETV2 DENGAN SEAT POSITION EMBEDDING. S1 thesis, Universitas Pendidikan Indonesia.
Abstract
Pemahaman terhadap arah perhatian visual siswa sangat penting dalam mengevaluasi keterlibatan mereka selama pembelajaran di kelas. Head Pose Estimation (HPE) menjadi metode yang efektif dalam mengidentifikasi focus atensi, namun penerapannya di kelas nyata sering terkendala oleh kualitas citra rendah dan posisi duduk siswa yang bervariasi yang menyebabkan metode regresi untuk memprediksi landmark wajah atau sudut Euler tidak optimal. Penelitian ini menggunakan pendekatan klasifikasi berbasis citra sebagai alternatif dan mengusulkan modifikasi arsitektur EfficientNetV2-S dengan menambahkan Seat Position Embedding (SPE) sebagai konteks spasial untuk meningkatkan akurasi klasifikasi pose kepala siswa. Set data dikembangkan melalui rekaman langsung di kelas dan diproses menjadi 4.574 gambar pose kepala dengan lima label arah (atas, bawah, depan, kanan, kiri). Evaluasi dilakukan pada beberapa arsitektur CNN dengan dan tanpa SPE. Hasil menunjukkan bahwa penambahan SPE pada model yang diusulkan memperoleh akurasi sebesar 83,25%, melebihi akurasi model baseline pada 82,53%. Pendekatan ini terbukti efisien dalam mengurangi ambiguitas visual dan memberikan interpretasi lebih akurat terhadap atensi siswa. Understanding students’ visual attention direction is essential for evaluating their engagement during classroom learning. Head Pose Estimation (HPE) is an effective method for identifying attention focus, however, its application in real classroom settings is often hindered by low image quality and varied student seating positions, which makes regression-based methods for predicting facial landmarks or Euler angles suboptimal. This study adopts an image-based classification approach as an alternative and proposes a modification of the EfficientNetV2-S architecture by integrating Seat Position Embedding (SPE) as spatial context to improve the accuracy of head pose classification. The dataset was developed from direct classroom recordings and processed into 4,574 head pose images with five directional labels (up, down, front, right, left). Several CNN architectures were evaluated with and without SPE. The results show that the proposed model with SPE achieved an accuracy of 83.25%, surpassing the baseline model’s accuracy of 82.53%. This approach has proven effective in reducing visual ambiguity and providing a more accurate interpretation of students' attention.
![]() |
Text
S_KOM_2100702_Title.pdf Download (669kB) |
![]() |
Text
S_KOM_2100702_Chapter1.pdf Download (254kB) |
![]() |
Text
S_KOM_2100702_Chapter2.pdf Restricted to Staf Perpustakaan Download (1MB) | Request a copy |
![]() |
Text
S_KOM_2100702_Chapter3.pdf Download (341kB) |
![]() |
Text
S_KOM_2100702_Chapter4.pdf Restricted to Staf Perpustakaan Download (2MB) | Request a copy |
![]() |
Text
S_KOM_2100702_Chapter5.pdf Download (242kB) |
Item Type: | Thesis (S1) |
---|---|
Additional Information: | https://scholar.google.com/citations?user=fv6VxHkAAAAJ&hl=en ID SINTA Dosen Pembimbing: Munir: 5974517 Yaya Wihardi: 5994413 |
Uncontrolled Keywords: | Deteksi Atensi Siswa, EfficientNetV2, Estimasi Pose Kepala, Ruang Kelas, Siswa Classroom, EfficientNetV2, Head Pose Estimation, Student, Student Attention Detection |
Subjects: | L Education > L Education (General) Q Science > QA Mathematics > QA75 Electronic computers. Computer science |
Divisions: | Fakultas Pendidikan Matematika dan Ilmu Pengetahuan Alam > Program Studi Ilmu Komputer |
Depositing User: | Ananda Myzza Marhelio |
Date Deposited: | 08 Sep 2025 04:40 |
Last Modified: | 08 Sep 2025 04:40 |
URI: | http://repository.upi.edu/id/eprint/137813 |
Actions (login required)
![]() |
View Item |