Universitas Stikubank (Unisbank) Semarang Repository

KLASIFIKASI JENIS TARI DAERAH BERDASARKAN POSE DENGAN CONVOLUTIONAL NEURAL NETWORKS (CNN)

SHUHADA, HASBI (2021) KLASIFIKASI JENIS TARI DAERAH BERDASARKAN POSE DENGAN CONVOLUTIONAL NEURAL NETWORKS (CNN). Undergraduate thesis, Universitas Stikubank.

[thumbnail of HLM JUDUL] PDF (HLM JUDUL)
Download (658kB)
[thumbnail of ABSTRAK] PDF (ABSTRAK)
Download (22kB)
[thumbnail of BAB I] PDF (BAB I)
Download (51kB)
[thumbnail of BAB II] PDF (BAB II)
Restricted to Repository staff only

Download (608kB)
[thumbnail of BAB III] PDF (BAB III)
Restricted to Repository staff only

Download (626kB)
[thumbnail of BAB IV] PDF (BAB IV)
Restricted to Repository staff only

Download (1MB)
[thumbnail of BAB V] PDF (BAB V)
Restricted to Repository staff only

Download (11kB)
[thumbnail of DAFTAR PUSTAKA] PDF (DAFTAR PUSTAKA)
Download (128kB)
[thumbnail of LAMPIRAN] PDF (LAMPIRAN)
Restricted to Repository staff only

Download (1MB)

Abstract

Seni tari daerah merupakan salah satu manifestasi dari keragaman budaya di Indonesia yang diciptakan dan berkembang secara turun temurun di suatu daerah tertentu. Perkembangan teknologi dan informasi saat ini tidak dapat dipungkiri berlangsung sangat cepat. Hal ini menyebabkan sebagian besar masyarakat khususnya di kalangan pemuda tidak mengenal tari tradisional yang ada di daerahnya sendiri. Menyikapi hal tersebut, penelitian ini dilakukan bertujuan untuk melakukan deteksi dan klasifikasi tari daerah berdasarkan pose. Metode yang digunakan adalah Convolutional Neural Network (CNN) dengan arsitektur BlazePose untuk deteksi pose dan algoritma Random Forest, Logistic Regression, dan K-Nearest Neighbour (KNN) untuk klasifikasi pose tari. Hasil pengujian dengan data test masing-masing algoritma memberikan akurasi yang signifikan seperti Random Forest 0.982, Logistic Regression 0.913, dan KNN 0.978. Berdasarkan hasil pengujian, Random Forest dipilih untuk pengujian dengan data nyata. Data yang digunkana 60 gambar tari dengan 20 gambar setiap kelas tari, proses pengujian menghasilkan nilai akurasi yang cukup baik yaitu 70%. Hasil penelitian ini dapat membantu masyarakat untuk mengenali jenis tari berdasarkan pose. Regional dance is one of the manifestations of cultural diversity in Indonesia which is created and developed from generation to generation in their area. The development of technology and information today is undeniably significant. This causes most of the community, especially among the youth, to be unfamiliar with traditional dances that exist in their area. In response to this, this research aims to detect and classify regional dances based on poses. The method used is Convolutional Neural Network (CNN) with BlazePose architecture for pose detection and Random Forest, Logistic Regression, and K-Nearest Neighbor (KNN) algorithms for dance pose classification. The test results with test data for each algorithm provide significant accuracies such as Random Forest 0.982, Logistic Regression 0.913, and KNN 0.978. Based on the test results, Random Forest was chosen for testing with actual data. The data used are 60 dance images with 20 images for each dance class. The Testing process produces a pretty good accuracy value of 70%. The results of this study can be used to help the public identify types of dance based on poses.

Item Type: Thesis (Undergraduate)
Additional Information: SKR.I.05.01.2028 NIM 17.01.53.0098
Uncontrolled Keywords: Klasifikasi Jenis Tari Daerah, Convolutional Neural Network, Deep Learning, Supervised Learning Classification of Regional Dance Types, Convolutional Neural Network, Deep Learning, Supervised Learning.
Subjects: Q Science > QA Mathematics > QA76 Computer software
Faculty / Institution: Fakultas Teknologi Informasi > Program Studi Teknik Informatika
Depositing User: Teteh Hayati
Date Deposited: 10 Nov 2021 03:21
Last Modified: 10 Nov 2021 03:21
URI: https://eprints.unisbank.ac.id/id/eprint/8024

Actions (login required)

View Item View Item