Please use this identifier to cite or link to this item:
https://ptsldigital.ukm.my/jspui/handle/123456789/486975Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor.advisor | Aini Hussain, Prof. Dr. | |
| dc.contributor.author | Aquache Mustapha (P37165) | |
| dc.date.accessioned | 2023-10-11T02:27:07Z | - |
| dc.date.available | 2023-10-11T02:27:07Z | - |
| dc.date.issued | 2012-07-11 | |
| dc.identifier.other | ukmvital:120122 | |
| dc.identifier.uri | https://ptsldigital.ukm.my/jspui/handle/123456789/486975 | - |
| dc.description | Advances in medical imaging technology enable rapid medical diagnoses with visualization and quantitative assessment by medical practitioners. Hence, technology has currently been adopted and used as a major tool in image-based medical diagnoses, such as in X-rays, MRI, ultrasounds, and so on. Therefore, the main objective of this study is to develop a computerized system to facilitate the assessment of vertebral fractures in making a spinal diagnosis. The development of three main modules, namely, indexing, classification, and retrieval modules is required to achieve this objective. The indexing module involves pre-processing to enhance and segment cervical and lumbar X-ray images. The adaptive factor based on the contrast adjustment technique was used for image enhancement and noise reduction. On the other hand, the active shape model (ASM) with two different shape boundary representations, known as 9-anatomical point representation (9-APR) and B-spline representation (B-SR), was used for vertebrae segmentation. ASM is an algorithm that is employed to extract unique and accurate features of vertebrae models for fracture characterization (FC) and ground truth establishment. Only two main FC algorithms are considered. This limits the problem to a two-categorical problem, that is, the region-based fracture characterization (RB-FC) and contour-based fracture characterization (CB-FC) categories. In RB-FC, all pixels within a shape are taken into account in extracting feature vectors. Moreover, the four algorithms that were considered are Gabor wavelets (GW), Gray level co-occurrence metric (GLCM), Radon transform (RT) and Orientation histogram (OH). For CB-FC, only shape boundary information is used such as the global shape profile (GSP) and shape signatures (SS). GSP, which includes measures of area, center of mass, perimeter, and so on, can only discriminate shapes with a large dissimilarity. Alternatively, SS contains information that represents shape features such as complex coordinates, and centroid distance. These features are sensitive to noise and not robust. As a result, spectral descriptors such as Fourier transform, which is simple to compute and robust for shape characterization, are also used together with the two previous representations. As for the testing and validation purposes of the classification module, Multi-layer perceptron using ten-fold cross validation (MLP-TFCV), nearest neighbor based vocabulary tree (NN-VT) and Fuzzy decision tree (FDT) are implemented to test the effectiveness of the features produced from the RB-FC and CB-FC techniques. During testing, the receiver operating characteristic (ROC) curve and area under curve (AUC) are used as statistical measurements in evaluating classifier performance. Next, vertebrae images with/without fractures are presented to the retrieval module wherein images are measured and ranked based on the minimum distance. Retrieval performance is evaluated using average retrieval rate (ARR) measure. Then, all three modules are integrated to produce a computerized system named as AVFAS, which is short for Assisted Vertebral Fracture Assessment System. The developed system shows promising results with a capability of indexing, classifying (>85%), and retrieving (>75%) vertebral fractures of spine images. In conclusion, this research has achieved its stated goal of developing an AVFAS for spinal/vertebrae fracture diagnosis that provides visualization and quantitative assessment to facilitate medical practitioners in carrying out rapid diagnoses.,Certification of Master's / Doctoral Thesis" is not available | |
| dc.language.iso | eng | |
| dc.publisher | UKM, Bangi | |
| dc.relation | Faculty of Engineering and Built Environment / Fakulti Kejuruteraan dan Alam Bina | |
| dc.rights | UKM | |
| dc.subject | Imaging systems in medicine | |
| dc.subject | Diagnostic imaging -- Digital techniques | |
| dc.subject | Diagnosis | |
| dc.subject | Computer-Assisted | |
| dc.subject | Computer vision in medicine | |
| dc.subject | Universiti Kebangsaan Malaysia -- Dissertations | |
| dc.subject | Dissertations, Academic -- Malaysia | |
| dc.title | Development of a computer assisted vertebral fracture assessment system (AVFAS) for vertebrae irregularity classification | |
| dc.type | Theses | |
| dc.format.pages | 293 | |
| dc.identifier.callno | R857.O6M848 2012 3 tesis | |
| dc.identifier.barcode | 004063(2019) | |
| Appears in Collections: | Faculty of Engineering and Built Environment / Fakulti Kejuruteraan dan Alam Bina | |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| ukmvital_120122+SOURCE1+SOURCE1.0.PDF Restricted Access | 734.7 kB | Adobe PDF |  View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.