Deteksi emosi wajah pelanggan setelah belanja melalui cctv toko dengan menggunakan haar cascade, kcf tracker, dan cnn

Hanitio, Edrick Hans (2025) Deteksi emosi wajah pelanggan setelah belanja melalui cctv toko dengan menggunakan haar cascade, kcf tracker, dan cnn. Bachelor thesis, Universitas Pelita Harapan.

[thumbnail of Title] Text (Title)
TITLE.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (196kB)
[thumbnail of Abstract] Text (Abstract)
abstract.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (455kB)
[thumbnail of Chapter 1] Text (Chapter 1)
CHAPTER 1.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (682kB)
[thumbnail of Chapter 2] Text (Chapter 2)
CHAPTER 2.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (2MB)
[thumbnail of Chapter 3] Text (Chapter 3)
CHAPTER 3.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (6MB)
[thumbnail of Chapter 4] Text (Chapter 4)
CHAPTER 4.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (1MB)
[thumbnail of Chapter 5] Text (Chapter 5)
CHAPTER 5.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (497kB)
[thumbnail of Bibliography] Text (Bibliography)
BIBLIOGRAPHY.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (369kB)
[thumbnail of Appendices] Text (Appendices)
APPENDICES.pdf
Restricted to Repository staff only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (5MB)

Abstract

Penelitian ini bertujuan untuk mengembangkan sebuah sistem deteksi emosi pelanggan secara real-time dengan memanfaatkan rekaman CCTV di lingkungan toko ritel. Sistem yang dirancang mengintegrasikan tiga metode utama, yaitu Haar Cascade untuk proses deteksi wajah, Kernelized Correlation Filters (KCF) Tracker untuk pelacakan wajah yang telah terdeteksi, serta Convolutional Neural Network (CNN) untuk klasifikasi ekspresi wajah ke dalam beberapa kategori emosi, meliputi senang, marah, sedih, netral, takut, jijik, dan terkejut. Pendekatan terintegrasi ini diharapkan dapat menghasilkan solusi yang efisien, praktis, serta mampu meningkatkan akurasi dalam analisis perilaku pelanggan secara otomatis. Metodologi penelitian melibatkan tahapan pengumpulan dataset ekspresi wajah, pelatihan model CNN dengan menggunakan data gambar maupun file CSV, serta pengujian sistem dalam berbagai kondisi nyata seperti variasi pencahayaan dan sudut pengambilan gambar. Hasil eksperimen menunjukkan bahwa Haar Cascade mampu mencapai tingkat akurasi hingga 99% dalam kondisi ideal, sementara KCF Tracker menunjukkan performa pelacakan yang lebih stabil pada objek wajah dengan pergerakan lambat. Selain itu, model CNN yang dilatih dengan data berformat CSV memperlihatkan peningkatan akurasi dibandingkan model yang dilatih menggunakan data gambar. Secara keseluruhan, hasil penelitian ini menunjukkan bahwa integrasi ketiga metode tersebut menghasilkan sistem deteksi dan klasifikasi emosi yang cukup akurat, yang berpotensi digunakan sebagai alat bantu dalam analisis perilaku pelanggan untuk mendukung pengambilan keputusan strategis pemilik toko, khususnya dalam upaya meningkatkan pengalaman berbelanja pelanggan melalui pemanfaatan teknologi berbasis kecerdasan buatan. / This study aims to develop a real-time customer emotion detection system utilizing CCTV footage in a retail store environment. The proposed system integrates three main methods, namely Haar Cascade for facial detection, Kernelized Correlation Filters (KCF) Tracker for tracking detected faces, and Convolutional Neural Network (CNN) for classifying facial expressions into several emotion categories, including happiness, anger, sadness, neutrality, fear, disgust, and surprise. This integrated approach is expected to provide an efficient, practical solution capable of enhancing the accuracy of automated customer behavior analysis. The research methodology involves collecting a facial expression dataset, training the CNN model using both image data and CSV files, and testing the system under various real-world scenarios, such as different lighting conditions and camera angles. Experimental results show that Haar Cascade achieves an accuracy rate of up to 99% under ideal conditions, while the KCF Tracker demonstrates more stable performance when tracking faces with slow movements. Moreover, the CNN model trained with CSV data exhibits improved accuracy compared to training based on image data. Overall, the results indicate that the integration of these three methods produces a sufficiently accurate system for detecting and classifying customer emotions, which can be utilized as a decision support tool for store owners, particularly in enhancing the shopping experience through the adoption of artificial intelligence-based technology.
Item Type: Thesis (Bachelor)
Creators:
Creators
NIM
Email
ORCID
Hanitio, Edrick Hans
NIM03082210038
edrickhans@gmail.com
UNSPECIFIED
Contributors:
Contribution
Contributors
NIDN/NIDK
Email
Thesis advisor
Robin, Robin
NIDN0116128001
UNSPECIFIED
Uncontrolled Keywords: Haar Cascade; KCF Tracker; CNN; Deteksi Emosi; CCTV; Ekspresi Wajah
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Depositing User: Edrick Hans Hanitio
Date Deposited: 21 Jul 2025 07:10
Last Modified: 21 Jul 2025 07:10
URI: http://repository.uph.edu/id/eprint/69809

Actions (login required)

View Item
View Item