DANCE PHRASE MOTION RECOGNITION USING LAYERED HIDDEN MARKOV MODEL DESIGN: CASE OF SIGEH PENGUTEN DANCE
Dance is formed by sequences of phrases performed on a time span. Phrases is constituted from gestures which is a sequence of poses. The hidden Markov model (HMM) in designed to recognize dance phrases in a layered scheme started from pose recognition. Key poses are identified by using K-means clust...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/36763 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Dance is formed by sequences of phrases performed on a time span. Phrases is constituted from gestures which is a sequence of poses. The hidden Markov model (HMM) in designed to recognize dance phrases in a layered scheme started from pose recognition. Key poses are identified by using K-means clustering method. Pose recognition then is done by HMM and compared with kNN (k - Nearest Neighbor). The pose sequence become an input to phrase recognition of Sigeh Penguten dance. The system is tested to recognize seven phrases on the Sigeh Penguten dance.
Phrases recognition using layered HMM scheme can recognize the seven phrases with 74% accuracy. Foot feature addition made distinguishing Samber Melayang and Kenui Melayang phrases achievable and the accuracy did get better. Pose recognition is more suitable using kNN with better performance for up to 20% against HMM. Implementation of k-Nearest Neighbor at pose layer increased the accuracy to 80%. Usage of kNN on pose layer is more effective than hidden Markov model. |
|---|