Classification of Indian Classical Dance Forms
Using Deep Learning
DEPARTMENT OF COMPUTER SCIENCE &
ENGINEERING
Under the Esteemed Guidance of Presented By
Mr. K.S.R PRASAD [Link]., A. GREESHMA (18MU1A0502)
Assistant Professor M. VASAVI (189P1A0546)
M. NAVYA MANI (189P1A0551)
S. PANDU RANGARAO (189P1A0534
ABSTRACT
Indian classical dance is the combination of gesture of all
the body parts. It has varied forms and is generally a combination
of single hand mudra, double hand mudra, leg alignment, hip
movement, eye movement, facial expression, and leg posture.
Each dance form has unique gesture, using which, they can be
classified. The costumes worn by dancers are also unique. This
work proposes the identification and classification of
Indian Classical Dance images using Deep Learning Convolution
Neural Network (CNN). This work uses the dataset consisting of
eight dance classes namely Bharatanatyam, Odissi, Kathak.
PROBLEM STTEMENT
Indian classical dance forms are practiced for 5000
years worldwide. However, it is difficult for a dance lover
to fully hold the content of the performance as it is made
up of hand poses, body poses, leg movements, hands
with respect to face and torso, and finally facial
expressions. All these movements should synchronize in
precision with both vocal song and the corresponding
music for various instruments
LITERATURE SURVEY
1. Novel feature extraction, selection and fusion for effective
malware family classification
Author: Mansour Ahmadi, Dmitry Ulyanov
Modern malware is designed with mutation
characteristics, namely polymorphism and metamorphism,
which causes an enormous growth in the number of variants of
malware samples. Categorization of malware samples on the
basis of their behaviors is essential for the computer security
community.
2. Efficient string matching: an aid to bibliographic
search. Communications of the ACM
Author: Alfred V Aho, Margaret J Corasick.
This paper describes a simple, efficient algorithm to
locate all occurrences of any of a finite number of
keywords in a string of text. The algorithm consists of
constructing a finite state pattern matching machine from
the keywords and then using the pattern matching machine
to process the text string in a single pass.
3. A meta-learning approach to automatic kernel
selection for support vector machines
Author: Shawkat Ali and Kate A Smith-Miles.
Appropriate choice of a kernel is the most important
ingredient of the kernel-based learning methods such as
support vector machine (SVM). Automatic kernel selection is a
key issue given the number of kernels available, and the
current trial- and-error nature of selecting the best kernel for a
given problem. This paper introduces a new method for
automatic kernel selection, with empirical results based on
classification.
4. Automatic model selection for the optimization of support
vector machine kernels
Author: Nedjem-Eddine Ayat, Mohamed Cheriet
This approach aims to optimize the kernel parameters and
to efficiently reduce the number of support vectors, so that the
generalization error can be reduced drastically. The proposed
methodology suggests the use of a new model selection criterion
based on the estimation of the probability of error of the SVM
classifier.
EXISTING SYSTEM
As rapid growth happening in Technology and
culture lot of people forgetting the culture in this Indian classical
dance is one category .Even by looking at the dance also now a
days lot of people unable to judge which kind of dance it is
exactly. To predict the kind of dance we are not having
any model as of now to overcome this problem in this project we
are proposing a powerful neural network model to detect
classical dance category.
DISADVANTAGES OF EXISTING
SYSTEM
It used to recognize only 5 types of dance forms.
Dataset is very less.
Model Accuracy is also just 75.83%.
PROPOSED SYSTEM
Indian Classical Dance Dataset used in this proposed model
has eight dance forms each consisting of 600 images with image
size of 224*224 in jpg format. The pre processed images are
normalized using resnet34 and are later trained using CNN model.
CNN is the backbone of this model with fully connected head and a
single hidden layer. We are performing image augmentation on the
data and applying appropriate pre processing techniques .After
refining the data for this problem statement we are applying
Inception V3 deep leaning CNN Architecture to train the model
and we are observing training and test accuracies of the model after
that we are checking performance of the model by plotting
confusion matrix and finally we designed a model in such a way
that if we give any indian classical dance image it will predict the
type of dance.
ADVANTAGES OF PROPOSED SYSTEM
It uses to recognize 8 types of dance forms.
Any type of Indian classical dance is predicted by this model.
Model accuracy is increased to 90%.
HARDWARE REQUIREMENTS
Processor : Core I3
RAM : 4GB(min)
Hard disk : 100GB
SOFTWARE REQUIREMENTS
Operating system : Windows 7 Ultimate.
Coding Language : Python
Front – End : Python
Data Base : ML and Deep Learning Model
THANK YOU