Microscopic Image Based Classification Algorithms for few Herbal Plants Identification using Machine Learning Approaches (Record no. 116711)

MARC details
000 -LEADER
fixed length control field ngm a22 7a 4500
008 - FIXED-LENGTH DATA ELEMENTS--GENERAL INFORMATION
fixed length control field 190506b ||||| |||| 00| 0 eng d
082 ## - DEWEY DECIMAL CLASSIFICATION NUMBER
Classification number TT000076
Item number FAT
100 ## - MAIN ENTRY--PERSONAL NAME
Personal name Fataniya, Bhupendra Damjibhai
245 ## - TITLE STATEMENT
Title Microscopic Image Based Classification Algorithms for few Herbal Plants Identification using Machine Learning Approaches
Statement of responsibility, etc by Bhupendra Damjibhai Fataniya
260 ## - PUBLICATION, DISTRIBUTION, ETC. (IMPRINT)
Place of publication, distribution, etc Ahmedabad
Name of publisher, distributor, etc Nirma Institute of Technology
Date of publication, distribution, etc 2018
300 ## - PHYSICAL DESCRIPTION
Extent 144p Ph. D. Thesis with Synopsis and CD.
500 ## - GENERAL NOTE
General note Guided by: Dr. Taniz Zaveri and Dr. Sanjeev Acharya With Synopsis and CD 11EXTPHDE76<br/><br/>ABSTRACT:<br/>Identification of herbal plant is of great interest in image processing and computer vision. In<br/>literature, many methods are presented to identify the herbal plant like leaf-based<br/>identification, chemical-based evaluation, physical evaluation and biological evaluation.<br/>Identification of herbal plant is more difficult and challenging when it is presented in powder<br/>form.<br/>This thesis presents microscopic image-based classification of a few herbal plants from its<br/>powder using various machine learning approaches. In this work, the cell characteristics of<br/>the herbal plants are studied. The dataset of the powder microscopic images of the three<br/>herbal plant is created in our laboratory using a Lawrence & Mayo microscope.<br/>It is found from the literature survey that various object can be uniquely represented by shape<br/>and texture based features which are further used for the object classification. In this thesis,<br/>shape and texture feature based novel methods for classification of herbal plants (Liquorice,<br/>Rhubarb and Dhatura) are proposed. The effectiveness of shape and texture feature methods<br/>are evaluated using different classifiers for classification. Three shape and five texture<br/>features are computed from the microscopic image dataset of the herbal plants. The<br/>effectiveness of the shape and texture based feature set and their combinations are<br/>investigated using Support Vector Machine (SVM), K-Nearest Neighbour (KNN) and<br/>Ensemble classifier. The effect of Speeded-Up Robust Features (SURF) is also analyzed<br/>using a different kernel of the SVM classifier. From the experiments highest classification<br/>accuracy of 99.9% is achieved when all the shape features with a combination of Gabor<br/>wavelet features are applied to quadratic SVM classifier. Finally, from the proposed<br/>algorithm, it is observed that the combination of selected shape and texture features work<br/>better for the classification of powder microscopic image of the herbal plant of Liquorice,<br/>Rhubarb and Dhatura. Convolutional Neural Networks (CNN) based classification is also carried out using pretrained<br/>CNN models like AlexNet, GoogLeNet and VGG for the herbal plants. Transfer<br/>learning approach is applied to train the pre-trained CNN models for our dataset. From the<br/>experiments, it is found that AlexNet-TL performs better and the results are compared with<br/>other pre-trained model GoogLeNet and VGG. It gives the highest accuracy of 99.49% with<br/>40x dataset when the single resolution dataset is considered for training CNN. In the case of<br/>the combination of the two datasets of different resolutions, VGG16-TL achieves the highest<br/>accuracy of 99.96% with 4x and 40x dataset. The combination of two or more datasets of<br/>different resolution increases the training complexity in the CNN model. GoogLeNet-TL<br/>performs better when all three resolution of the dataset are combined. It achieves the highest<br/>accuracy of 99.03%. The fourth category of the class where the background of the image has<br/>been included with the dataset to train the pre-trained CNN model. In this category of dataset,<br/>AlexNet-TL gives 98.65% accuracy which is better than other pre-trained CNN model. It is<br/>observed from the simulation results, transfer learning from the pre-trained CNN model<br/>works better for automatic classification and highest classification accuracy of 99.96% is<br/>achieved in VGG16-TL for the combination of 4x and 40x dataset. When dataset of 4x, 10x<br/>and 40x is applied to the CNN for training, transfer learning work well with 40x dataset. It is<br/>due to the clear visibility of cell characteristics at 40x resolution.<br/>The features extracted from the microscopic image of the herbal plant using shape and texture<br/>are hand crafted features. It is very difficult to tune hand crafted features with a classifier for<br/>the herbal plant classification. In the CNN based classification of herbal plants, features are<br/>self-extracted from the input image dataset. CNN based classification is the good choice<br/>when input image dataset is complex and difficult to tune the hand crafted features. It takes<br/>the more training time to train CNN model. This problem can be overcome by using the<br/>Graphic Processor Unit (GPU) during the training of CNN.
856 ## - ELECTRONIC LOCATION AND ACCESS
Uniform Resource Identifier http://repository.nirmauni.ac.in/jspui/handle/123456789/8385
Public note Institute Repository (Campus Access)
856 ## - ELECTRONIC LOCATION AND ACCESS
Uniform Resource Identifier https://shodhganga.inflibnet.ac.in/jspui/handle/10603/246013
Public note Shodhganga
942 ## - ADDED ENTRY ELEMENTS (KOHA)
Source of classification or shelving scheme Dewey Decimal Classification
Koha item type Thesis

No items available.

© 2025 by NIMA Knowledge Centre, Ahmedabad.
Koha version 24.05