MARC details
| 000 -LEADER |
|---|
| fixed length control field |
15862nam a22002657a 4500 |
| 008 - FIXED-LENGTH DATA ELEMENTS--GENERAL INFORMATION |
|---|
| fixed length control field |
201210s2024 a|||f bm|| 00| 0 eng d |
| 024 7# - Author Identifier |
|---|
| Standard number or code |
0000-0003-3060-5986 |
| Source of number or code |
ORCID |
| 040 ## - CATALOGING SOURCE |
|---|
| Original cataloging agency |
EG-CaNU |
| Transcribing agency |
EG-CaNU |
| 041 0# - Language Code |
|---|
| Language code of text |
eng |
| Language code of abstract |
eng |
| -- |
ara |
| 082 ## - DEWEY DECIMAL CLASSIFICATION NUMBER |
|---|
| Classification number |
610 |
| 100 0# - MAIN ENTRY--PERSONAL NAME |
|---|
| Personal name |
Samar Ibrahim Antar Awwad Ramadan |
| 245 1# - TITLE STATEMENT |
|---|
| Title |
Enhancing Lung Cancer Detection |
| Remainder of title |
: Leveraging CT Features for Identifying Nodules of Various Sizes in Chest X-Ray Images |
| Statement of responsibility, etc. |
/Samar Ibrahim Antar Awwad Ramadan |
| 260 ## - PUBLICATION, DISTRIBUTION, ETC. |
|---|
| Date of publication, distribution, etc. |
2024 |
| 300 ## - PHYSICAL DESCRIPTION |
|---|
| Extent |
125 p. |
| Other physical details |
ill. |
| Dimensions |
21 cm. |
| 500 ## - GENERAL NOTE |
|---|
| Materials specified |
Supervisor:<br/>Prof. Mustafa Elattar |
| 502 ## - Dissertation Note |
|---|
| Dissertation type |
Thesis (M.A.)—Nile University, Egypt, 2024 . |
| 504 ## - Bibliography |
|---|
| Bibliography |
"Includes bibliographical references" |
| 505 0# - Contents |
|---|
| Formatted contents note |
Contents:<br/>Table of Contents<br/>ACKNOWLEDGEMENT................................................................................................... i<br/>ABSTRACT........................................................................................................................ ii<br/>Table of Contents............................................................................................................... iii<br/>List of Figures................................................................................................................... vii<br/>List of Tables .......................................................................................................................x<br/>List of Equations................................................................................................................ xi<br/>List of Abbreviation......................................................................................................... xiii<br/>Chapter 1 INTRODUCTION...............................................................................................1<br/>1.1. MOTIVATION....................................................................................................1<br/>1.2. PROBLEM DEFINITION...................................................................................3<br/>1.3. CONTRIBUTION................................................................................................4<br/>1.4. THESIS OUTLINES ...........................................................................................5<br/>Chapter 2 BACKGROUND.................................................................................................7<br/>2.1. OVERVIEW ........................................................................................................7<br/>2.2. CLINICAL BACKGROUND..............................................................................8<br/>2.2.1. Respiratory System..........................................................................................8<br/>2.2.2. Lung Cancer...................................................................................................10<br/>2.2.3. Nodule Definitions and Different size of nodules .........................................10<br/>2.2.4. Types of Nodules (Benign vs malignant) ......................................................13<br/>2.2.5. Statistics of Lung Cancer Spread...................................................................15<br/>2.3. LUNG CANCER DIAGNOSIS.........................................................................16<br/>2.4. SUMMARY.......................................................................................................18<br/>Chapter 3 RELATED WORK ...........................................................................................19<br/>3.1. LUNG CANCER DETECTION & SEGMENTATION IN CT........................19<br/>3.2. LUNG CANCER DETECTION & SEGMENTATION IN X-ray....................25<br/>Chapter 4 DATA AND DATASET CREATION..............................................................31<br/>4.1. PUBLIC DATASETS........................................................................................31<br/>4.1.1. Chexpert.........................................................................................................31<br/>4.1.2. The Japanese Society of Radiological Technology .......................................31<br/>4.1.3. Montgomery (NLM-MC)...............................................................................32<br/>4.1.4. Covid-19 and Lung Opacity Datasets............................................................32<br/>iv<br/>4.1.5. LIDC-IDRI.....................................................................................................33<br/>4.1.6. Nodules Chest X-rays (LIDC-IDRI)..............................................................34<br/>4.2. SIMULATED DATASET .................................................................................34<br/>4.2.1. Simulated Data for Image Enhancement .......................................................34<br/>4.3. CREATE SYNTHESIZED 2D CXR FROM 3D THORAX CT.......................35<br/>4.4. ANNOTATION .................................................................................................38<br/>4.5. CORRECTION FOR BOUNDING BOX ANNOTATION AFTER <br/>AUGMENTATION........................................................................................................38<br/>4.6. SUMMARY.......................................................................................................40<br/>Chapter 5 PROPOSED FRAMEWORK ...........................................................................42<br/>5.1. OVERVIEW ......................................................................................................42<br/>5.2. IMAGE PRE-PROCESSING ............................................................................42<br/>5.2.1. Convert CXR images to grayscale and Crop to square..................................43<br/>5.2.2. Adaptive Histogram Equalization..................................................................44<br/>5.2.3. Augmentation.................................................................................................45<br/>5.2.4. Expanding Montgomery Dataset by Merging Lungs.....................................45<br/>5.2.5. The Laplacian of Gaussian (LoG)..................................................................46<br/>5.2.6. Gamma Correction.........................................................................................46<br/>5.2.7. CT Pre-processing..........................................................................................47<br/>5.3. IMAGE ENHANCEMENT...............................................................................49<br/>5.3.1. Proposed image enhancement approach VAE + U-Net:................................49<br/>5.4. LUNG SEGMENTATION................................................................................51<br/>5.5. BONE SUPPRESSION .....................................................................................53<br/>5.5.1. Segmenting Bones from DRRs Using U-Net ................................................53<br/>5.5.2. Generating Rib and Difference images for JSRT ..........................................55<br/>5.5.3. Proposed Latent Boosted Pix2Pix++ Architecture (LB-Pix2Pix++).............56<br/>5.6. CLASSIFICATION ...........................................................................................60<br/>5.7. EXPLAINABLE AI (XAI)................................................................................61<br/>5.7.1. Class Activation Mapping (CAM).................................................................63<br/>5.7.2. Gradient-Weighted Class Activation Mapping (Grad-CAM)........................64<br/>5.8. LOCALIZATION..............................................................................................67<br/>Chapter 6 EXPERIMENTAL SETTINGS ........................................................................68<br/>v<br/>6.1. IMAGE ENHANCEMENT...............................................................................68<br/>6.1.1. Overview........................................................................................................68<br/>6.1.2. VAE Objective Function Optimization and Training....................................68<br/>6.2. LUNG SEGMENTATION................................................................................69<br/>6.2.1. Overview........................................................................................................69<br/>6.2.2. Loss Functions...............................................................................................70<br/>6.2.3. Evaluation matrices........................................................................................71<br/>6.3. BONE SUPPRESSION .....................................................................................72<br/>6.3.1. Bone Segmentation Training .........................................................................72<br/>6.3.2. LB-Pix2Pix++ for Bone Suppression ............................................................72<br/>6.3.3. Evaluation Protocol........................................................................................74<br/>6.4. CLASIFICATION .............................................................................................75<br/>6.4.1. Without data augmentation............................................................................75<br/>6.4.2. With applying data augmentation ..................................................................76<br/>6.4.3. Generalization on JSRT data .........................................................................77<br/>6.5. LOCALIZATION..............................................................................................77<br/>6.5.1. Bounding box estimation (detection).............................................................77<br/>6.5.2. The segmentation model (2nd technique)......................................................78<br/>Chapter 7 RESULTS..........................................................................................................80<br/>7.1. IMAGE ENHANEEMENT ...............................................................................80<br/>7.1.1. Simulation of extra/ halting structures data ...................................................80<br/>7.1.2. The training outcome .....................................................................................80<br/>7.2. LUNG SEGMENTATION................................................................................82<br/>7.3. TESTING OF VAEU-NET AND RESU-NET++.............................................84<br/>7.4. BONE SUPPRESSION .....................................................................................85<br/>7.4.1. Ablation Study ...............................................................................................85<br/>7.4.2. Comparison to the State-of-the-Art ...............................................................89<br/>7.5. CLASSIFICATION ...........................................................................................90<br/>7.5.1. Without data augmentation............................................................................90<br/>7.5.2. With data augmentation .................................................................................90<br/>7.5.3. Generalization on JSRT data & Performance Comparison to Literature ......91<br/>7.5.4. Grad-CAM for the JSRT dataset....................................................................92<br/>vi<br/>7.5.5. Testing on unseen dataset ..............................................................................95<br/>7.6. LOCALIZATION..............................................................................................96<br/>7.6.1. Bounding box estimation (Detection) on JSRT.............................................96<br/>7.6.2. Testing Bounding box detection model on unseen datasets ..........................97<br/>7.6.3. The segmentation model (2nd technique) on JSRT.......................................99<br/>7.6.4. Testing the segmentation model on DRRs...................................................100<br/>Chapter 8 DISCUSSION .................................................................................................103<br/>8.1. LUNG SEGMENTATION..............................................................................103<br/>8.2. TESTING OF VAEU-NET AND RESU-NET++...........................................104<br/>8.3. BONE SUPPRESSION ...................................................................................104<br/>8.4. CLASSIFICATION .........................................................................................107<br/>8.5. LOCALIZATION............................................................................................108<br/>Chapter 9 CONCLUSION AND FUTURE WORK........................................................110<br/>9.1. CONCLUSION................................................................................................110<br/>9.2. FUTURE WORK.............................................................................................111<br/>REFERENCES ................................................................................................................112<br/>Appendix A EXPLAINABILITY....................................................................................121<br/>A.1. CLASS-ENHANCED ATTENTIVE RESPONSE (CLEAR)............................121<br/>A.2. DECONVOLUTIONAL NETWORKS .............................................................121<br/>A.3. LAYER-WISE RELEVANCE PROPAGATION (LRP)...................................123<br/>A.4. DIFFERENCE OF LRP FROM GRAD-CAM AND CAM...............................124<br/>A.5. LRP RULES FOR DEEP RECTIFIER NETWORKS .......................................124<br/>A.6. DEEP LEARNING IMPORTANT FEATURES (DEEPLIFT) .........................125 |
| 520 3# - Abstract |
|---|
| Abstract |
Abstract:<br/>The high-resolution capability of thoracic Computerized Tomography (CT) <br/>scans enables detailed visualization of lung structures, making it possible to <br/>detect even small abnormalities. This level of detail is especially crucial for <br/>identifying early-stage lung nodules, which may not be visible or <br/>distinguishable in other imaging modalities. However, CT scans have <br/>limitations like cost, access, radiation exposure, and lack of portability. Chest <br/>X-rays (CXRs) are 2D projective images where objects may not appear clear <br/>and can overlap, leading to limited visibility of lung nodules. Deep learning <br/>algorithms have the potential to address these limitations. This thesis proposes <br/>leveraging CT volume features to improve the detection and localization of <br/>lung cancer in CXR, thus enabling radiologists to diagnose lung cancer using <br/>Chest X-ray images more confidently. This could be done by finding CT <br/>features that define small, medium, and large-sized nodules and projecting <br/>those features onto CXR data. The proposed detection system includes all <br/>phases and subtasks required for CXR-based lung cancer identification. Due <br/>to a scarcity of CXR images that contain different nodule sizes, feasibility of <br/>small nodules detection in CXR images is demonstrated by fine-tuning our <br/>predictive model using Digitally Reconstructed Radiographs (DRRs), created <br/>from CT projections. Our proposed method outperforms previous state-of-theart methods, in terms of Mean Intersection Over Union (mIOU) and AUC for <br/>lung cancer detection. Furthermore, the proposed method incorporates <br/>Explainable AI (XAI) module to provide physicians with trustworthy <br/>decision-making tools. It achieves AUC of 0.9762 and MIOU of 0.9025 for <br/>detecting and segmenting nodules on the JSRT dataset, which contains <br/>different sizes of nodules. Additionally, it achieves MIOU of 0.6579 and <br/>0.7646 for small and medium size nodules, respectively.<br/>Keywords:<br/>Chest X-ray, Lung Cancer Segmentation, Computerized Tomography, Different Sizes of <br/>Nodules, Bone Suppression, Chest X-ray Enhancement and DRRs Generation. |
| 546 ## - Language Note |
|---|
| Language Note |
Text in English, abstracts in English and Arabic |
| 650 #4 - Subject |
|---|
| Subject |
informatics |
| 655 #7 - Index Term-Genre/Form |
|---|
| Source of term |
NULIB |
| focus term |
Dissertation, Academic |
| 690 ## - Subject |
|---|
| School |
informatics |
| 942 ## - ADDED ENTRY ELEMENTS (KOHA) |
|---|
| Source of classification or shelving scheme |
Dewey Decimal Classification |
| Koha item type |
Thesis |
| 655 #7 - Index Term-Genre/Form |
|---|
| -- |
187 |