Classifying Indian Classical Dances By Motion Posture Patterns
Dance is a classic form of human motion which is usually performed as a reaction of expression to music. The Indian classical dances, for instance, require multiple complicated movements that relates to body motion postures and hand gestures with high similarities. Past studies showed interests...
| Main Author: | |
|---|---|
| Format: | Monograph |
| Language: | English |
| Published: |
Universiti Sains Malaysia
2019
|
| Subjects: | |
| Online Access: | http://eprints.usm.my/58278/1/Classifying%20Indian%20Classical%20Dances%20By%20Motion%20Posture%20Patterns.pdf |
| _version_ | 1825907866371883008 |
|---|---|
| author | Khor, Chong Hwa |
| author_facet | Khor, Chong Hwa |
| author_sort | Khor, Chong Hwa |
| collection | USM |
| description | Dance is a classic form of human motion which is usually performed as a
reaction of expression to music. The Indian classical dances, for instance, require
multiple complicated movements that relates to body motion postures and hand gestures
with high similarities. Past studies showed interests using various methods to classify
dances. The most common method used is the Hidden Markov Models (HMM), apart
from using the correlation matrix method and hierarchical cluster analysis. Nevertheless,
less effort has been placed in analysing the Indian dance by using the data mining
approach. Therefore, the objectives in this work are to (i) distinguish different types of
Indian classical dances, (ii) classify the type of dance based on motion posture patterns
and (iii) determine the effects of attributes on the classification accuracy. This study
involves five types of Indian classical dances (Kathak, Bharatanatyam, Kuchipudi,
Manipuri and Odissi) motion postures. The data mining approaches were used to
classify the motion posture patterns by type of dances. A total of 15 dance videos were
collected from the public available domain for body joints tracking processes using the
Kinovea software. Data mining analysis was performed in three stages: data pre�processing, data classification and knowledge discovery using the WEKA software.
RandomForest algorithm returned the highest classification accuracy (99.2616%). On
attribute configuration, y-coordinates of left wrist (LW(y)) was identified as the most
significant attribute to differentiate the Indian classical dance classes. |
| first_indexed | 2024-03-06T16:10:28Z |
| format | Monograph |
| id | usm.eprints-58278 |
| institution | Universiti Sains Malaysia |
| language | English |
| last_indexed | 2024-03-06T16:10:28Z |
| publishDate | 2019 |
| publisher | Universiti Sains Malaysia |
| record_format | dspace |
| spelling | usm.eprints-582782023-04-28T09:57:12Z http://eprints.usm.my/58278/ Classifying Indian Classical Dances By Motion Posture Patterns Khor, Chong Hwa T Technology TJ Mechanical engineering and machinery Dance is a classic form of human motion which is usually performed as a reaction of expression to music. The Indian classical dances, for instance, require multiple complicated movements that relates to body motion postures and hand gestures with high similarities. Past studies showed interests using various methods to classify dances. The most common method used is the Hidden Markov Models (HMM), apart from using the correlation matrix method and hierarchical cluster analysis. Nevertheless, less effort has been placed in analysing the Indian dance by using the data mining approach. Therefore, the objectives in this work are to (i) distinguish different types of Indian classical dances, (ii) classify the type of dance based on motion posture patterns and (iii) determine the effects of attributes on the classification accuracy. This study involves five types of Indian classical dances (Kathak, Bharatanatyam, Kuchipudi, Manipuri and Odissi) motion postures. The data mining approaches were used to classify the motion posture patterns by type of dances. A total of 15 dance videos were collected from the public available domain for body joints tracking processes using the Kinovea software. Data mining analysis was performed in three stages: data pre�processing, data classification and knowledge discovery using the WEKA software. RandomForest algorithm returned the highest classification accuracy (99.2616%). On attribute configuration, y-coordinates of left wrist (LW(y)) was identified as the most significant attribute to differentiate the Indian classical dance classes. Universiti Sains Malaysia 2019-06-01 Monograph NonPeerReviewed application/pdf en http://eprints.usm.my/58278/1/Classifying%20Indian%20Classical%20Dances%20By%20Motion%20Posture%20Patterns.pdf Khor, Chong Hwa (2019) Classifying Indian Classical Dances By Motion Posture Patterns. Project Report. Universiti Sains Malaysia, Pusat Pengajian Kejuruteraan Mekanik. (Submitted) |
| spellingShingle | T Technology TJ Mechanical engineering and machinery Khor, Chong Hwa Classifying Indian Classical Dances By Motion Posture Patterns |
| title | Classifying Indian Classical Dances By Motion Posture Patterns |
| title_full | Classifying Indian Classical Dances By Motion Posture Patterns |
| title_fullStr | Classifying Indian Classical Dances By Motion Posture Patterns |
| title_full_unstemmed | Classifying Indian Classical Dances By Motion Posture Patterns |
| title_short | Classifying Indian Classical Dances By Motion Posture Patterns |
| title_sort | classifying indian classical dances by motion posture patterns |
| topic | T Technology TJ Mechanical engineering and machinery |
| url | http://eprints.usm.my/58278/1/Classifying%20Indian%20Classical%20Dances%20By%20Motion%20Posture%20Patterns.pdf |
| work_keys_str_mv | AT khorchonghwa classifyingindianclassicaldancesbymotionposturepatterns |