Pratama, Satrya Fajri (2018) Three-Dimensional Exact Legendre Moment Invariants For Amphetamine-Type Stimulants Molecular Structure Representation. Doctoral thesis, UTeM.
![[img]](http://eprints.utem.edu.my/style/images/fileicons/text.png) |
Text (24 Pages)
Three-Dimensional Exact Legendre Moment Invariants For Amphetamine-Type Stimulants Molecular Structure Representation.pdf - Submitted Version Download (432kB) |
|
Text (Full Text)
Three-Dimensional Exact Legendre Moment Invariants For Amphetamine-Type Stimulants Molecular Structure Representation.pdf - Submitted Version Restricted to Registered users only Download (3MB) |
Abstract
The abuse of amphetamine-type stimulants (ATS) drugs has become a global,harrowing social problem.The technical limitations of the current test kits to detect new brand of ATS drugs present a challenge to national law enforcement authorities and scientific staff of forensic laboratories.Meanwhile,new molecular imaging devices which allowed mankind to characterize the physical three-dimensional (3D) molecular structure have been recently introduced,and it can be used to remedy the limitations of existing drug test kits.Thus,a new type of 3D molecular structure representation technique,or molecular descriptors,should be developed to cater the 3D molecular structure acquired physically using these molecular imaging devices.One of the image processing methods to represent a 3D image is 3D moments and moment invariants. However,there are problems exhibited by the existing 3D moments and moment invariants.Therefore,it is necessary to propose a new 3D moment invariants which is free from these problems.This study compares various 3D moments and identified 3D Legendre moments as the best moments to construct 3D moment invariants,namely 3D exact Legendre moment invariants (3D ELMI),which is used to represent the 3D molecular structure of ATS drugs.Since the 3D molecular structure of ATS drugs dataset obtained using molecular imaging devices are currently unavailable,this study acquired the 3D molecular structure of ATS drugs data from United Nations Office of Drug and Crime (UNODC) and pihkal.info database instead.The proposed technique was compared to the existing 3D moment invariants and molecular descriptors techniques in terms of processing time,memory consumption,single instance invariance,intra- and inter-class variance,and classification accuracy.The comparative study conducted found that 3D ELMI performs better than the existing 3D moment invariants,such as 3D geometric moment invariants (3D GMI),3D Gaussian–Hermite moment invariants (3D GHMI),and 3D Zernike descriptors (3D ZD).The satisfactory performance of 3D ELMI is attributed to numerous factors,such as the quality of the 3D Legendre,exact computation of the 3D Legendre,and the novelty of the proposed invariants techniques.The proposed technique was also compared to existing 3D molecular descriptors,for example weighted holistic invariants molecular (WHIM),geometry,topology,and atom weights assembly (GETAWAY),radial distribution function (RDF),and 3D molecule representation of structure based on electron diffraction (3D-MoRSE) descriptors.Despite 3D ELMI is capable to overcome the limitations of existing 3D molecular descriptors which depends on 3D molecular structure model instead of physical molecular structure obtained from molecular imaging devices,the test reveals 3D ELMI is not as good as these techniques,primarily due to the substantial number of features produced by the proposed technique.Nevertheless,the promising applicability and the unique approach of the proposed technique to represent the 3D molecular structure of ATS drugs has been demonstrated and worth to receive further exploration in the future works.
| Item Type: | Thesis (Doctoral) |
|---|---|
| Uncontrolled Keywords: | Image processing, Pattern recognition systems, Drugs, Analysis, Automation, Three-Dimensional Exact Legendre Moment Invariants, Amphetamine-Type Stimulants Molecular Structure Representation |
| Subjects: | T Technology > T Technology (General) T Technology > TA Engineering (General). Civil engineering (General) |
| Divisions: | Library > Tesis > FTMK |
| Depositing User: | Mohd. Nazir Taib |
| Date Deposited: | 04 Sep 2019 08:21 |
| Last Modified: | 30 Mar 2022 08:39 |
| URI: | http://eprints.utem.edu.my/id/eprint/23366 |
| Statistic Details: | View Download Statistic |
Actions (login required)
 |
View Item |