SlideShare a Scribd company logo
1 of 13
Download to read offline
Contrast and Resolution Improvement of
POCUS Using Self-Consistent CycleGAN
Shujaat Khan, Jaeyoung Huh, and Jong Chul Ye
Department of Bio and Brain Engineering, Korea Advanced Institute of Science
and Technology (KAIST), 335 Gwahangno, Yuseong-gu, Daejeon 305-701, Korea.
{shujaat,woori93,jong.ye}@kaist.ac.kr
https://bispl.weebly.com/
In conjunction with
Contents
I. Introduction
I. Motivation
II. Problem formulation
II. Proposed method
I. Divide and conquer
II. Self-consistent cycleGAN
III. Network architecture
IV. Dataset
III. Results
I. Qualitative results
II. Quantitative results
III. Comparative results
IV. Conclusion
2
Applications Needs
 Efficacy
 Portability
 Low Cost
 Accessibility
Possibilities
Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 3
Introduction: motivation
• Can save the treatment cost about 33% by reducing the referral
to secondary care.
• Expected to become an in-demand electric gadget for personal
and tele-health monitoring
low-contrast/low-resolution high-contrast/high-resolution
Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 4
Introduction: problem formulation
low-contrast/low-resolution high-contrast/high-resolution
high-contrast/low-resolution
Contrast
enhancement
Network
Super-resolution
Network
Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 5
Proposed method: divide and conquer
- fixed weights
high-contrast/high-resolution
high-contrast/low-resolution
Unsupervised training Supervised training
:high-contrast low-resolution
:low-contrast/ low-resolution
Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 6
Proposed method: self-consistent cycleGAN
Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 7
Proposed method: network architecture
The proposed method was implemented using Python with TensorFlow on an Nvidia GEFORCE GTX 1080 Ti GPU.
For network training Adam optimizer was used and learning rate linearly changed from 5 × 10−4
→ 1 × 10−4
in 200 epochs.
Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 8
Proposed method: Dataset
LOW-QUALITY INPUT IMAGES HIGH-QUALITY TARGET IMAGES
Scanner
NPUS050 portable US system.
NOPROBLEM MEDICAL Co., China
L3–12 linear array,
E-CUBE 12R US Alpinion Co., Korea
Description
• 200 in vivo (carotid/thyroid regions of
2 healthy volunteers)
• 200 ATS-539 phantom images
• 320 in vivo (carotid/thyroid regions of 8
healthy volunteers)
• 192 ATS-539 phantom images.
• Filtered by NLLR[1] and DeepDeconv [2]
Training • 157 in vivo and 168 phantom All
Testing • 43 in vivo and 32 phantom None
[1] Zhu, Lei, et al, IEEE CVPR 2017 [2] Khan et al, IEEE TUFFC, 2021
Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 9
Results: qualitative evaluation
Reconstruction results on in vivo and phantom data
Sample
results
from
different
regions
of
phantom
and
anatomical
regions.
Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 10
Results: quantitative evaluation
• For the simulated low quality images that are generated from 2×
axially and 2× laterally sub-sampled images of high-end images,
the PSNR and SSIM metrics show 13.58 dB and 0.63 units gain.
• On average, the proposed method achieves 14.96 dB, 2.38, 0.86
04 units CR, CNR and GCNR, which is 21.77%, 30.06%, and 44.4
2% higher, respectively, than those of the input images.
• On average the reconstruction time for a single image is around
13.18 (milliseconds), and it could further reduce by optimized i
mplementation.
Axial
depth(mm)
36
27
18
9
0
45
Phantom
in
vivo
Lateral length(mm)
0 20
10 30 38.2
(a) Input image (b) histogram equalization (c) standard CycleGAN (d) our method
A B
C
Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 11
Results: comparative evaluation
Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 12
Conclusion
• The proposed method can directly process unpaired image domai
n data to generate high-quality noise and artifact free images from
low-quality images.
• In contrast to the existing approaches, we decomposed the contra
st enhancement and resolution improvement into two steps and sh
owed improved performance without spurious artifacts.
• Even though proposed method shows noticeable gain in image qua
lity, the results are still not ideal and require a comprehensive clini
cal comparison.
• However, experimental results on in vivo and phantom data sugges
t that the proposed schemes may substantially help in designing lo
w-powered, high quality pocus systems.
Thank You

More Related Content

Similar to Contrast and resolution improvement of pocus using self consistent cycle gan

Image Reconstruction in Computed Tomography
Image Reconstruction in Computed TomographyImage Reconstruction in Computed Tomography
Image Reconstruction in Computed TomographyAnjan Dangal
 
A Review of Image Contrast Enhancement Techniques
A Review of Image Contrast Enhancement TechniquesA Review of Image Contrast Enhancement Techniques
A Review of Image Contrast Enhancement TechniquesIRJET Journal
 
Lung Tumour Detection using Image Processing
Lung Tumour Detection using Image Processing Lung Tumour Detection using Image Processing
Lung Tumour Detection using Image Processing Aviral Chaurasia
 
Unsupervised Deconvolution Neural Network for High Quality Ultrasound Imaging
Unsupervised Deconvolution Neural Network for High Quality Ultrasound ImagingUnsupervised Deconvolution Neural Network for High Quality Ultrasound Imaging
Unsupervised Deconvolution Neural Network for High Quality Ultrasound ImagingShujaat Khan
 
Image Contrast Enhancement Approach using Differential Evolution and Particle...
Image Contrast Enhancement Approach using Differential Evolution and Particle...Image Contrast Enhancement Approach using Differential Evolution and Particle...
Image Contrast Enhancement Approach using Differential Evolution and Particle...IRJET Journal
 
Deep Generative model-based quality control for cardiac MRI segmentation
Deep Generative model-based quality control for cardiac MRI segmentation Deep Generative model-based quality control for cardiac MRI segmentation
Deep Generative model-based quality control for cardiac MRI segmentation Seunghyun Hwang
 
Novel algorithm for color image demosaikcing using laplacian mask
Novel algorithm for color image demosaikcing using laplacian maskNovel algorithm for color image demosaikcing using laplacian mask
Novel algorithm for color image demosaikcing using laplacian maskeSAT Journals
 
Inference of Nonlinear Gene Regulatory Networks through Optimized Ensemble of...
Inference of Nonlinear Gene Regulatory Networks through Optimized Ensemble of...Inference of Nonlinear Gene Regulatory Networks through Optimized Ensemble of...
Inference of Nonlinear Gene Regulatory Networks through Optimized Ensemble of...Arinze Akutekwe
 
A DISCUSSION ON IMAGE ENHANCEMENT USING HISTOGRAM EQUALIZATION BY VARIOUS MET...
A DISCUSSION ON IMAGE ENHANCEMENT USING HISTOGRAM EQUALIZATION BY VARIOUS MET...A DISCUSSION ON IMAGE ENHANCEMENT USING HISTOGRAM EQUALIZATION BY VARIOUS MET...
A DISCUSSION ON IMAGE ENHANCEMENT USING HISTOGRAM EQUALIZATION BY VARIOUS MET...pharmaindexing
 
Deep Conditional Adversarial learning for polyp Segmentation
Deep Conditional Adversarial learning for polyp SegmentationDeep Conditional Adversarial learning for polyp Segmentation
Deep Conditional Adversarial learning for polyp Segmentationmultimediaeval
 
MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...
MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...
MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...ijma
 
Dual energy CT in radiotherapy: Current applications and future outlook
Dual energy CT in radiotherapy: Current applications and future outlookDual energy CT in radiotherapy: Current applications and future outlook
Dual energy CT in radiotherapy: Current applications and future outlookWookjin Choi
 
Single image super resolution with improved wavelet interpolation and iterati...
Single image super resolution with improved wavelet interpolation and iterati...Single image super resolution with improved wavelet interpolation and iterati...
Single image super resolution with improved wavelet interpolation and iterati...iosrjce
 
Shape and Level Bottles Detection Using Local Standard Deviation and Hough Tr...
Shape and Level Bottles Detection Using Local Standard Deviation and Hough Tr...Shape and Level Bottles Detection Using Local Standard Deviation and Hough Tr...
Shape and Level Bottles Detection Using Local Standard Deviation and Hough Tr...IJECEIAES
 
Saliency Based Hookworm and Infection Detection for Wireless Capsule Endoscop...
Saliency Based Hookworm and Infection Detection for Wireless Capsule Endoscop...Saliency Based Hookworm and Infection Detection for Wireless Capsule Endoscop...
Saliency Based Hookworm and Infection Detection for Wireless Capsule Endoscop...IRJET Journal
 
Cycle-free CycleGAN using invertible generator for unsupervised low-dose CT d...
Cycle-free CycleGAN using invertible generator for unsupervised low-dose CT d...Cycle-free CycleGAN using invertible generator for unsupervised low-dose CT d...
Cycle-free CycleGAN using invertible generator for unsupervised low-dose CT d...KwonTaesung
 
IMAGE RESOLUTION ENHANCEMENT BY USING SWT AND DWT
IMAGE RESOLUTION ENHANCEMENT BY USING SWT AND DWTIMAGE RESOLUTION ENHANCEMENT BY USING SWT AND DWT
IMAGE RESOLUTION ENHANCEMENT BY USING SWT AND DWTIRJET Journal
 

Similar to Contrast and resolution improvement of pocus using self consistent cycle gan (20)

Image Reconstruction in Computed Tomography
Image Reconstruction in Computed TomographyImage Reconstruction in Computed Tomography
Image Reconstruction in Computed Tomography
 
A Review of Image Contrast Enhancement Techniques
A Review of Image Contrast Enhancement TechniquesA Review of Image Contrast Enhancement Techniques
A Review of Image Contrast Enhancement Techniques
 
Lung Tumour Detection using Image Processing
Lung Tumour Detection using Image Processing Lung Tumour Detection using Image Processing
Lung Tumour Detection using Image Processing
 
Unsupervised Deconvolution Neural Network for High Quality Ultrasound Imaging
Unsupervised Deconvolution Neural Network for High Quality Ultrasound ImagingUnsupervised Deconvolution Neural Network for High Quality Ultrasound Imaging
Unsupervised Deconvolution Neural Network for High Quality Ultrasound Imaging
 
Image Contrast Enhancement Approach using Differential Evolution and Particle...
Image Contrast Enhancement Approach using Differential Evolution and Particle...Image Contrast Enhancement Approach using Differential Evolution and Particle...
Image Contrast Enhancement Approach using Differential Evolution and Particle...
 
Deep Generative model-based quality control for cardiac MRI segmentation
Deep Generative model-based quality control for cardiac MRI segmentation Deep Generative model-based quality control for cardiac MRI segmentation
Deep Generative model-based quality control for cardiac MRI segmentation
 
Novel algorithm for color image demosaikcing using laplacian mask
Novel algorithm for color image demosaikcing using laplacian maskNovel algorithm for color image demosaikcing using laplacian mask
Novel algorithm for color image demosaikcing using laplacian mask
 
Inference of Nonlinear Gene Regulatory Networks through Optimized Ensemble of...
Inference of Nonlinear Gene Regulatory Networks through Optimized Ensemble of...Inference of Nonlinear Gene Regulatory Networks through Optimized Ensemble of...
Inference of Nonlinear Gene Regulatory Networks through Optimized Ensemble of...
 
A DISCUSSION ON IMAGE ENHANCEMENT USING HISTOGRAM EQUALIZATION BY VARIOUS MET...
A DISCUSSION ON IMAGE ENHANCEMENT USING HISTOGRAM EQUALIZATION BY VARIOUS MET...A DISCUSSION ON IMAGE ENHANCEMENT USING HISTOGRAM EQUALIZATION BY VARIOUS MET...
A DISCUSSION ON IMAGE ENHANCEMENT USING HISTOGRAM EQUALIZATION BY VARIOUS MET...
 
Deep Conditional Adversarial learning for polyp Segmentation
Deep Conditional Adversarial learning for polyp SegmentationDeep Conditional Adversarial learning for polyp Segmentation
Deep Conditional Adversarial learning for polyp Segmentation
 
MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...
MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...
MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...
 
Dual energy CT in radiotherapy: Current applications and future outlook
Dual energy CT in radiotherapy: Current applications and future outlookDual energy CT in radiotherapy: Current applications and future outlook
Dual energy CT in radiotherapy: Current applications and future outlook
 
Single image super resolution with improved wavelet interpolation and iterati...
Single image super resolution with improved wavelet interpolation and iterati...Single image super resolution with improved wavelet interpolation and iterati...
Single image super resolution with improved wavelet interpolation and iterati...
 
F010224446
F010224446F010224446
F010224446
 
Viva201393(1).pptxbaru
Viva201393(1).pptxbaruViva201393(1).pptxbaru
Viva201393(1).pptxbaru
 
Shape and Level Bottles Detection Using Local Standard Deviation and Hough Tr...
Shape and Level Bottles Detection Using Local Standard Deviation and Hough Tr...Shape and Level Bottles Detection Using Local Standard Deviation and Hough Tr...
Shape and Level Bottles Detection Using Local Standard Deviation and Hough Tr...
 
Saliency Based Hookworm and Infection Detection for Wireless Capsule Endoscop...
Saliency Based Hookworm and Infection Detection for Wireless Capsule Endoscop...Saliency Based Hookworm and Infection Detection for Wireless Capsule Endoscop...
Saliency Based Hookworm and Infection Detection for Wireless Capsule Endoscop...
 
Cycle-free CycleGAN using invertible generator for unsupervised low-dose CT d...
Cycle-free CycleGAN using invertible generator for unsupervised low-dose CT d...Cycle-free CycleGAN using invertible generator for unsupervised low-dose CT d...
Cycle-free CycleGAN using invertible generator for unsupervised low-dose CT d...
 
Review DRCN
Review DRCNReview DRCN
Review DRCN
 
IMAGE RESOLUTION ENHANCEMENT BY USING SWT AND DWT
IMAGE RESOLUTION ENHANCEMENT BY USING SWT AND DWTIMAGE RESOLUTION ENHANCEMENT BY USING SWT AND DWT
IMAGE RESOLUTION ENHANCEMENT BY USING SWT AND DWT
 

More from Shujaat Khan

Physics informed deep learning for efficient b-mode ultrasound imaging
Physics informed deep learning for efficient b-mode ultrasound imagingPhysics informed deep learning for efficient b-mode ultrasound imaging
Physics informed deep learning for efficient b-mode ultrasound imagingShujaat Khan
 
Variational formulation of unsupervised deep learning for ultrasound image ar...
Variational formulation of unsupervised deep learning for ultrasound image ar...Variational formulation of unsupervised deep learning for ultrasound image ar...
Variational formulation of unsupervised deep learning for ultrasound image ar...Shujaat Khan
 
Universal plane wave compounding for high quality us imaging using deep learning
Universal plane wave compounding for high quality us imaging using deep learningUniversal plane wave compounding for high quality us imaging using deep learning
Universal plane wave compounding for high quality us imaging using deep learningShujaat Khan
 
Switchable and tunable deep beamformer using adaptive instance normalization ...
Switchable and tunable deep beamformer using adaptive instance normalization ...Switchable and tunable deep beamformer using adaptive instance normalization ...
Switchable and tunable deep beamformer using adaptive instance normalization ...Shujaat Khan
 
Clutter suppressed deep beamformer for echocardiography using deep learning
Clutter suppressed deep beamformer for echocardiography using deep learningClutter suppressed deep beamformer for echocardiography using deep learning
Clutter suppressed deep beamformer for echocardiography using deep learningShujaat Khan
 
Adaptive and compressive beamforming using deep learning for medical ultrasound
Adaptive and compressive beamforming using deep learning for medical ultrasoundAdaptive and compressive beamforming using deep learning for medical ultrasound
Adaptive and compressive beamforming using deep learning for medical ultrasoundShujaat Khan
 
Deep Learning-Based Universal Beamformer for Ultrasound Imaging
Deep Learning-Based Universal Beamformer for Ultrasound ImagingDeep Learning-Based Universal Beamformer for Ultrasound Imaging
Deep Learning-Based Universal Beamformer for Ultrasound ImagingShujaat Khan
 
Switchable Deep Beamformer for Ultrasound Imaging Using ADAIN
Switchable Deep Beamformer for Ultrasound Imaging Using ADAINSwitchable Deep Beamformer for Ultrasound Imaging Using ADAIN
Switchable Deep Beamformer for Ultrasound Imaging Using ADAINShujaat Khan
 

More from Shujaat Khan (8)

Physics informed deep learning for efficient b-mode ultrasound imaging
Physics informed deep learning for efficient b-mode ultrasound imagingPhysics informed deep learning for efficient b-mode ultrasound imaging
Physics informed deep learning for efficient b-mode ultrasound imaging
 
Variational formulation of unsupervised deep learning for ultrasound image ar...
Variational formulation of unsupervised deep learning for ultrasound image ar...Variational formulation of unsupervised deep learning for ultrasound image ar...
Variational formulation of unsupervised deep learning for ultrasound image ar...
 
Universal plane wave compounding for high quality us imaging using deep learning
Universal plane wave compounding for high quality us imaging using deep learningUniversal plane wave compounding for high quality us imaging using deep learning
Universal plane wave compounding for high quality us imaging using deep learning
 
Switchable and tunable deep beamformer using adaptive instance normalization ...
Switchable and tunable deep beamformer using adaptive instance normalization ...Switchable and tunable deep beamformer using adaptive instance normalization ...
Switchable and tunable deep beamformer using adaptive instance normalization ...
 
Clutter suppressed deep beamformer for echocardiography using deep learning
Clutter suppressed deep beamformer for echocardiography using deep learningClutter suppressed deep beamformer for echocardiography using deep learning
Clutter suppressed deep beamformer for echocardiography using deep learning
 
Adaptive and compressive beamforming using deep learning for medical ultrasound
Adaptive and compressive beamforming using deep learning for medical ultrasoundAdaptive and compressive beamforming using deep learning for medical ultrasound
Adaptive and compressive beamforming using deep learning for medical ultrasound
 
Deep Learning-Based Universal Beamformer for Ultrasound Imaging
Deep Learning-Based Universal Beamformer for Ultrasound ImagingDeep Learning-Based Universal Beamformer for Ultrasound Imaging
Deep Learning-Based Universal Beamformer for Ultrasound Imaging
 
Switchable Deep Beamformer for Ultrasound Imaging Using ADAIN
Switchable Deep Beamformer for Ultrasound Imaging Using ADAINSwitchable Deep Beamformer for Ultrasound Imaging Using ADAIN
Switchable Deep Beamformer for Ultrasound Imaging Using ADAIN
 

Recently uploaded

Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsJoaquim Jorge
 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)Gabriella Davis
 
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationFrom Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationSafe Software
 
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...apidays
 
Automating Google Workspace (GWS) & more with Apps Script
Automating Google Workspace (GWS) & more with Apps ScriptAutomating Google Workspace (GWS) & more with Apps Script
Automating Google Workspace (GWS) & more with Apps Scriptwesley chun
 
Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobe
Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, AdobeApidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobe
Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobeapidays
 
Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024The Digital Insurer
 
MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024MIND CTI
 
Partners Life - Insurer Innovation Award 2024
Partners Life - Insurer Innovation Award 2024Partners Life - Insurer Innovation Award 2024
Partners Life - Insurer Innovation Award 2024The Digital Insurer
 
Strategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherStrategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherRemote DBA Services
 
HTML Injection Attacks: Impact and Mitigation Strategies
HTML Injection Attacks: Impact and Mitigation StrategiesHTML Injection Attacks: Impact and Mitigation Strategies
HTML Injection Attacks: Impact and Mitigation StrategiesBoston Institute of Analytics
 
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Drew Madelung
 
🐬 The future of MySQL is Postgres 🐘
🐬  The future of MySQL is Postgres   🐘🐬  The future of MySQL is Postgres   🐘
🐬 The future of MySQL is Postgres 🐘RTylerCroy
 
A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?Igalia
 
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...Martijn de Jong
 
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024Rafal Los
 
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024The Digital Insurer
 
Manulife - Insurer Innovation Award 2024
Manulife - Insurer Innovation Award 2024Manulife - Insurer Innovation Award 2024
Manulife - Insurer Innovation Award 2024The Digital Insurer
 
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers:  A Deep Dive into Serverless Spatial Data and FMECloud Frontiers:  A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FMESafe Software
 
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...Miguel Araújo
 

Recently uploaded (20)

Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and Myths
 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)
 
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationFrom Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
 
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
 
Automating Google Workspace (GWS) & more with Apps Script
Automating Google Workspace (GWS) & more with Apps ScriptAutomating Google Workspace (GWS) & more with Apps Script
Automating Google Workspace (GWS) & more with Apps Script
 
Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobe
Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, AdobeApidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobe
Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobe
 
Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024
 
MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024
 
Partners Life - Insurer Innovation Award 2024
Partners Life - Insurer Innovation Award 2024Partners Life - Insurer Innovation Award 2024
Partners Life - Insurer Innovation Award 2024
 
Strategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherStrategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a Fresher
 
HTML Injection Attacks: Impact and Mitigation Strategies
HTML Injection Attacks: Impact and Mitigation StrategiesHTML Injection Attacks: Impact and Mitigation Strategies
HTML Injection Attacks: Impact and Mitigation Strategies
 
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
 
🐬 The future of MySQL is Postgres 🐘
🐬  The future of MySQL is Postgres   🐘🐬  The future of MySQL is Postgres   🐘
🐬 The future of MySQL is Postgres 🐘
 
A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?
 
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...
 
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024
 
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
 
Manulife - Insurer Innovation Award 2024
Manulife - Insurer Innovation Award 2024Manulife - Insurer Innovation Award 2024
Manulife - Insurer Innovation Award 2024
 
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers:  A Deep Dive into Serverless Spatial Data and FMECloud Frontiers:  A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
 
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
 

Contrast and resolution improvement of pocus using self consistent cycle gan

  • 1. Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN Shujaat Khan, Jaeyoung Huh, and Jong Chul Ye Department of Bio and Brain Engineering, Korea Advanced Institute of Science and Technology (KAIST), 335 Gwahangno, Yuseong-gu, Daejeon 305-701, Korea. {shujaat,woori93,jong.ye}@kaist.ac.kr https://bispl.weebly.com/ In conjunction with
  • 2. Contents I. Introduction I. Motivation II. Problem formulation II. Proposed method I. Divide and conquer II. Self-consistent cycleGAN III. Network architecture IV. Dataset III. Results I. Qualitative results II. Quantitative results III. Comparative results IV. Conclusion 2
  • 3. Applications Needs  Efficacy  Portability  Low Cost  Accessibility Possibilities Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 3 Introduction: motivation • Can save the treatment cost about 33% by reducing the referral to secondary care. • Expected to become an in-demand electric gadget for personal and tele-health monitoring
  • 4. low-contrast/low-resolution high-contrast/high-resolution Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 4 Introduction: problem formulation
  • 5. low-contrast/low-resolution high-contrast/high-resolution high-contrast/low-resolution Contrast enhancement Network Super-resolution Network Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 5 Proposed method: divide and conquer
  • 6. - fixed weights high-contrast/high-resolution high-contrast/low-resolution Unsupervised training Supervised training :high-contrast low-resolution :low-contrast/ low-resolution Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 6 Proposed method: self-consistent cycleGAN
  • 7. Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 7 Proposed method: network architecture The proposed method was implemented using Python with TensorFlow on an Nvidia GEFORCE GTX 1080 Ti GPU. For network training Adam optimizer was used and learning rate linearly changed from 5 × 10−4 → 1 × 10−4 in 200 epochs.
  • 8. Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 8 Proposed method: Dataset LOW-QUALITY INPUT IMAGES HIGH-QUALITY TARGET IMAGES Scanner NPUS050 portable US system. NOPROBLEM MEDICAL Co., China L3–12 linear array, E-CUBE 12R US Alpinion Co., Korea Description • 200 in vivo (carotid/thyroid regions of 2 healthy volunteers) • 200 ATS-539 phantom images • 320 in vivo (carotid/thyroid regions of 8 healthy volunteers) • 192 ATS-539 phantom images. • Filtered by NLLR[1] and DeepDeconv [2] Training • 157 in vivo and 168 phantom All Testing • 43 in vivo and 32 phantom None [1] Zhu, Lei, et al, IEEE CVPR 2017 [2] Khan et al, IEEE TUFFC, 2021
  • 9. Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 9 Results: qualitative evaluation Reconstruction results on in vivo and phantom data Sample results from different regions of phantom and anatomical regions.
  • 10. Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 10 Results: quantitative evaluation • For the simulated low quality images that are generated from 2× axially and 2× laterally sub-sampled images of high-end images, the PSNR and SSIM metrics show 13.58 dB and 0.63 units gain. • On average, the proposed method achieves 14.96 dB, 2.38, 0.86 04 units CR, CNR and GCNR, which is 21.77%, 30.06%, and 44.4 2% higher, respectively, than those of the input images. • On average the reconstruction time for a single image is around 13.18 (milliseconds), and it could further reduce by optimized i mplementation.
  • 11. Axial depth(mm) 36 27 18 9 0 45 Phantom in vivo Lateral length(mm) 0 20 10 30 38.2 (a) Input image (b) histogram equalization (c) standard CycleGAN (d) our method A B C Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 11 Results: comparative evaluation
  • 12. Contrast and Resolution Improvement of POCUS Using Self-Consistent CycleGAN 12 Conclusion • The proposed method can directly process unpaired image domai n data to generate high-quality noise and artifact free images from low-quality images. • In contrast to the existing approaches, we decomposed the contra st enhancement and resolution improvement into two steps and sh owed improved performance without spurious artifacts. • Even though proposed method shows noticeable gain in image qua lity, the results are still not ideal and require a comprehensive clini cal comparison. • However, experimental results on in vivo and phantom data sugges t that the proposed schemes may substantially help in designing lo w-powered, high quality pocus systems.