SlideShare a Scribd company logo
1 of 3
Download to read offline
Client
A company developing a healthcare platform hired Elinext to help it build a pneumonia
diagnosis tool.
Challenge
The company has been developing a comprehensive healthcare platform. Treating pneumonia
has been one of its focus areas due to COVID, and it wanted to build a pneumonia diagnosis
tool for the platform.
The tool was destined to analyze lung X-ray images and identify signs of pneumonia using
machine learning (ML), an artificial intelligence (AI) technique. The company didn’t have
relevant in-house experts, so they reached out for help and found it with Elinext.
Solution
We began by looking for a neural network that would best analyze lung images and found four
candidates: ResNet (50, 101, 152), VGG (16, 19), MobileNet and Inception (V2, V3). After
digging deeper into each of them, we chose InceptionV3 developed by Google Research Lab.
Once we chose our neural network, we moved on to designing the software architecture and
training the algorithm.
Architecture
The software is based on web technology and can be integrated into other systems like desktop
applications and mobile apps.
We used publicly available frameworks, libraries and technologies to develop the software. To
create a static HTML5 web page, we deployed a web server in a Docker container. On that
page, a user can upload a lung image and get feedback. The image is sent for processing
through the HTTP protocol.
Training
Training is the most challenging part in building ML algorithms. Your ability to source enough
data, avoid errors and be consistent throughout the process can make or break the algorithm.
Manual training is often inconsistent. You may forget which steps you have taken and in which
order, or occasionally delete logs. As a result, you won’t be able to accurately repeat a training
session. Therefore, we automated the process from A to Z.
We needed to train complex models with huge datasets fast. To do that, we rented an Amazon
Web Services (AWS) g3s.xlarge instance and used Deep Learning Base AMI (Ubuntu 18.10). The
latter is a powerful machine boasting 16GB of RAM, a 4-core CPU and an Nvidia Tesla M60 GPU.
It was a perfect fit for the task. Once we have chosen the technology, the training could begin.
We built a clean Docker container to isolate the model from outer influences and downloaded a
ton of lung images from Kaggle. To be able to work with the images, we subsampled them,
narrowing them down to a relevant and consistent selection. The dataset and training
environment were ready.
The training began. We faced a challenge in overtraining, whereby the model could memorize
training images and as a result fail to accurately analyze new images in the future. Our solution
was to slightly modify the images’ width, height, graininess and some other parameters. We
also launched Tensorboard to monitor training metrics.
At the final stages, we exported the model to an H5 file, a format commonly used across
industries from healthcare to aerospace, for testing. We tested it manually and automatically,
using preset scripts.
Accuracy
The model we’ve developed has a margin of confidence and uses binary identification. What
does this mean? It means if the algorithm identifies 80% of lungs as unaffected, it will say the
lungs are healthy. If the figure is below 80%, it will assume the lungs might be affected and
require medical attention.
How It Works
The user opens the web application in their browser, uploads a lung image, sends it to the
service and receives feedback. The feedback will show whether the lungs are healthy or if a
doctor should take a look at the image.
Result
The tool we’ve built can help reduce human error in identifying pneumonia. This is particularly
useful during the pandemic when doctors are overloaded and might overlook some signs of
illness. We can also scale the model up to identify some other diseases. Scaling the model down
will help integrate it into other systems, speed things up and allow for the analysis of multiple
images simultaneously.
Pneumonia diagnosis tool Case Study

More Related Content

Similar to Pneumonia diagnosis tool Case Study

XYZ Fast Prototyping MGMT 3405 1 Definition – Fa.docx
XYZ Fast Prototyping MGMT 3405  1  Definition – Fa.docxXYZ Fast Prototyping MGMT 3405  1  Definition – Fa.docx
XYZ Fast Prototyping MGMT 3405 1 Definition – Fa.docx
jeffevans62972
 
Focus magazine cloud article
Focus magazine cloud articleFocus magazine cloud article
Focus magazine cloud article
Khashi Rahmani
 

Similar to Pneumonia diagnosis tool Case Study (20)

The Evolution Of Eclipse 1. 1 )
The Evolution Of Eclipse 1. 1 )The Evolution Of Eclipse 1. 1 )
The Evolution Of Eclipse 1. 1 )
 
ML Times: Mainframe Machine Learning Initiative- June newsletter (2018)
ML Times: Mainframe Machine Learning Initiative- June newsletter (2018)ML Times: Mainframe Machine Learning Initiative- June newsletter (2018)
ML Times: Mainframe Machine Learning Initiative- June newsletter (2018)
 
SAP Development Object Testing
SAP Development Object TestingSAP Development Object Testing
SAP Development Object Testing
 
AI NOTES.docx
AI NOTES.docxAI NOTES.docx
AI NOTES.docx
 
Machine Learning in Malware Detection
Machine Learning in Malware DetectionMachine Learning in Malware Detection
Machine Learning in Malware Detection
 
Final .pptx
Final .pptxFinal .pptx
Final .pptx
 
Machine Learning in Static Analysis of Program Source Code
Machine Learning in Static Analysis of Program Source CodeMachine Learning in Static Analysis of Program Source Code
Machine Learning in Static Analysis of Program Source Code
 
Top Artificial Intelligence Tools & Frameworks in 2023.pdf
Top Artificial Intelligence Tools & Frameworks in 2023.pdfTop Artificial Intelligence Tools & Frameworks in 2023.pdf
Top Artificial Intelligence Tools & Frameworks in 2023.pdf
 
9 Digital Transformation Trends for 2023.pdf
9 Digital Transformation Trends for 2023.pdf9 Digital Transformation Trends for 2023.pdf
9 Digital Transformation Trends for 2023.pdf
 
1645 goldenberg using our laptop
1645 goldenberg using our laptop1645 goldenberg using our laptop
1645 goldenberg using our laptop
 
Bringing Machine Learning to Mobile Apps with TensorFlow
Bringing Machine Learning to Mobile Apps with TensorFlowBringing Machine Learning to Mobile Apps with TensorFlow
Bringing Machine Learning to Mobile Apps with TensorFlow
 
implementing_ai_for_improved_performance_testing_the_key_to_success.pptx
implementing_ai_for_improved_performance_testing_the_key_to_success.pptximplementing_ai_for_improved_performance_testing_the_key_to_success.pptx
implementing_ai_for_improved_performance_testing_the_key_to_success.pptx
 
Human Emotion Recognition using Machine Learning
Human Emotion Recognition using Machine LearningHuman Emotion Recognition using Machine Learning
Human Emotion Recognition using Machine Learning
 
XYZ Fast Prototyping MGMT 3405 1 Definition – Fa.docx
XYZ Fast Prototyping MGMT 3405  1  Definition – Fa.docxXYZ Fast Prototyping MGMT 3405  1  Definition – Fa.docx
XYZ Fast Prototyping MGMT 3405 1 Definition – Fa.docx
 
MongoDB World 2018: Building Intelligent Apps with MongoDB & Google Cloud
MongoDB World 2018: Building Intelligent Apps with MongoDB & Google CloudMongoDB World 2018: Building Intelligent Apps with MongoDB & Google Cloud
MongoDB World 2018: Building Intelligent Apps with MongoDB & Google Cloud
 
RSA 2015 Blending the Automated and the Manual: Making Application Vulnerabil...
RSA 2015 Blending the Automated and the Manual: Making Application Vulnerabil...RSA 2015 Blending the Automated and the Manual: Making Application Vulnerabil...
RSA 2015 Blending the Automated and the Manual: Making Application Vulnerabil...
 
Focus magazine cloud article
Focus magazine cloud articleFocus magazine cloud article
Focus magazine cloud article
 
UNCOVERING FAKE NEWS BY MEANS OF SOCIAL NETWORK ANALYSIS
UNCOVERING FAKE NEWS BY MEANS OF SOCIAL NETWORK ANALYSISUNCOVERING FAKE NEWS BY MEANS OF SOCIAL NETWORK ANALYSIS
UNCOVERING FAKE NEWS BY MEANS OF SOCIAL NETWORK ANALYSIS
 
UNCOVERING FAKE NEWS BY MEANS OF SOCIAL NETWORK ANALYSIS
UNCOVERING FAKE NEWS BY MEANS OF SOCIAL NETWORK ANALYSISUNCOVERING FAKE NEWS BY MEANS OF SOCIAL NETWORK ANALYSIS
UNCOVERING FAKE NEWS BY MEANS OF SOCIAL NETWORK ANALYSIS
 
A-Hospital-Management-System Shanto , waliul , Turjo , Munna- FULL update 2 ...
A-Hospital-Management-System Shanto  , waliul , Turjo , Munna- FULL update 2 ...A-Hospital-Management-System Shanto  , waliul , Turjo , Munna- FULL update 2 ...
A-Hospital-Management-System Shanto , waliul , Turjo , Munna- FULL update 2 ...
 

More from Elinext

More from Elinext (14)

Modular Web Design.pdf
Modular Web Design.pdfModular Web Design.pdf
Modular Web Design.pdf
 
Data Migration Testing Purpose, Test Strategy And Scenarios.pdf
Data Migration Testing Purpose, Test Strategy And Scenarios.pdfData Migration Testing Purpose, Test Strategy And Scenarios.pdf
Data Migration Testing Purpose, Test Strategy And Scenarios.pdf
 
Building a social network website from scratch
Building a social network website from scratchBuilding a social network website from scratch
Building a social network website from scratch
 
Software Testing QA: Automated Testing vs. Manual Testing. Which to Use, and ...
Software Testing QA: Automated Testing vs. Manual Testing. Which to Use, and ...Software Testing QA: Automated Testing vs. Manual Testing. Which to Use, and ...
Software Testing QA: Automated Testing vs. Manual Testing. Which to Use, and ...
 
Development Standards and Regulations for HealthTech
Development Standards and Regulations for HealthTechDevelopment Standards and Regulations for HealthTech
Development Standards and Regulations for HealthTech
 
Virtual Clinics In the USA
Virtual Clinics In the USAVirtual Clinics In the USA
Virtual Clinics In the USA
 
Notifications in Health Apps
Notifications in Health AppsNotifications in Health Apps
Notifications in Health Apps
 
Сomparison table of culture parameters for major outsourcing countries
Сomparison table of culture parameters for major outsourcing countriesСomparison table of culture parameters for major outsourcing countries
Сomparison table of culture parameters for major outsourcing countries
 
History and Trends of FinTech in Germany, Austria and Switzerland
History and Trends of FinTech in Germany, Austria and SwitzerlandHistory and Trends of FinTech in Germany, Austria and Switzerland
History and Trends of FinTech in Germany, Austria and Switzerland
 
Develpment of an electronics website Case Study
Develpment of an electronics website Case StudyDevelpment of an electronics website Case Study
Develpment of an electronics website Case Study
 
Case Study_Employee skill assessment software
Case Study_Employee skill assessment softwareCase Study_Employee skill assessment software
Case Study_Employee skill assessment software
 
Case Study_Application for integration with financial organizations
Case Study_Application for integration with financial organizationsCase Study_Application for integration with financial organizations
Case Study_Application for integration with financial organizations
 
Elinext Healthcare Software Development
Elinext Healthcare Software DevelopmentElinext Healthcare Software Development
Elinext Healthcare Software Development
 
The Digitalization of European SMEs
The Digitalization of European SMEsThe Digitalization of European SMEs
The Digitalization of European SMEs
 

Recently uploaded

Real Sex Provide In Goa ✂️ Call Girl (9316020077) Call Girl In Goa
Real Sex Provide In Goa ✂️ Call Girl   (9316020077) Call Girl In GoaReal Sex Provide In Goa ✂️ Call Girl   (9316020077) Call Girl In Goa
Real Sex Provide In Goa ✂️ Call Girl (9316020077) Call Girl In Goa
Real Sex Provide In Goa
 
Abortion pills Buy Farwaniya (+918133066128) Cytotec 200mg tablets Al AHMEDI
Abortion pills Buy Farwaniya (+918133066128) Cytotec 200mg tablets Al AHMEDIAbortion pills Buy Farwaniya (+918133066128) Cytotec 200mg tablets Al AHMEDI
Abortion pills Buy Farwaniya (+918133066128) Cytotec 200mg tablets Al AHMEDI
Abortion pills in Kuwait Cytotec pills in Kuwait
 
@Safe Abortion pills IN Jeddah(+918133066128) Un_wanted kit Buy Jeddah
@Safe Abortion pills IN Jeddah(+918133066128) Un_wanted kit Buy Jeddah@Safe Abortion pills IN Jeddah(+918133066128) Un_wanted kit Buy Jeddah
@Safe Abortion pills IN Jeddah(+918133066128) Un_wanted kit Buy Jeddah
Abortion pills in Kuwait Cytotec pills in Kuwait
 
No Advance 931~602~0077 Goa ✂️ Call Girl , Indian Call Girl Goa For Full nig...
No Advance  931~602~0077 Goa ✂️ Call Girl , Indian Call Girl Goa For Full nig...No Advance  931~602~0077 Goa ✂️ Call Girl , Indian Call Girl Goa For Full nig...
No Advance 931~602~0077 Goa ✂️ Call Girl , Indian Call Girl Goa For Full nig...
Real Sex Provide In Goa
 
❤️ Chandigarh Call Girls ☎️99158-51334☎️ Escort service in Chandigarh ☎️ Chan...
❤️ Chandigarh Call Girls ☎️99158-51334☎️ Escort service in Chandigarh ☎️ Chan...❤️ Chandigarh Call Girls ☎️99158-51334☎️ Escort service in Chandigarh ☎️ Chan...
❤️ Chandigarh Call Girls ☎️99158-51334☎️ Escort service in Chandigarh ☎️ Chan...
rajveerescorts2022
 
Obat Penggugur Kandungan Cytotec Dan Gastrul Harga Indomaret
Obat Penggugur Kandungan Cytotec Dan Gastrul Harga IndomaretObat Penggugur Kandungan Cytotec Dan Gastrul Harga Indomaret
Obat Penggugur Kandungan Cytotec Dan Gastrul Harga Indomaret
Cara Menggugurkan Kandungan 087776558899
 
Spauldings classification ppt by Dr C P PRINCE
Spauldings classification ppt by Dr C P PRINCESpauldings classification ppt by Dr C P PRINCE
Spauldings classification ppt by Dr C P PRINCE
DR.PRINCE C P
 
Goa Call Girl 931~602~0077 Call ✂️ Girl Service Vip Top Model Safe
Goa Call Girl  931~602~0077 Call ✂️ Girl Service Vip Top Model SafeGoa Call Girl  931~602~0077 Call ✂️ Girl Service Vip Top Model Safe
Goa Call Girl 931~602~0077 Call ✂️ Girl Service Vip Top Model Safe
Real Sex Provide In Goa
 

Recently uploaded (20)

MAGNESIUM - ELECTROLYTE IMBALANCE (HYPERMAGNESEMIA & HYPOMAGNESEMIA).pdf
MAGNESIUM - ELECTROLYTE IMBALANCE (HYPERMAGNESEMIA & HYPOMAGNESEMIA).pdfMAGNESIUM - ELECTROLYTE IMBALANCE (HYPERMAGNESEMIA & HYPOMAGNESEMIA).pdf
MAGNESIUM - ELECTROLYTE IMBALANCE (HYPERMAGNESEMIA & HYPOMAGNESEMIA).pdf
 
Test Bank -Medical-Surgical Nursing Concepts for Interprofessional Collaborat...
Test Bank -Medical-Surgical Nursing Concepts for Interprofessional Collaborat...Test Bank -Medical-Surgical Nursing Concepts for Interprofessional Collaborat...
Test Bank -Medical-Surgical Nursing Concepts for Interprofessional Collaborat...
 
Real Sex Provide In Goa ✂️ Call Girl (9316020077) Call Girl In Goa
Real Sex Provide In Goa ✂️ Call Girl   (9316020077) Call Girl In GoaReal Sex Provide In Goa ✂️ Call Girl   (9316020077) Call Girl In Goa
Real Sex Provide In Goa ✂️ Call Girl (9316020077) Call Girl In Goa
 
Abortion pills Buy Farwaniya (+918133066128) Cytotec 200mg tablets Al AHMEDI
Abortion pills Buy Farwaniya (+918133066128) Cytotec 200mg tablets Al AHMEDIAbortion pills Buy Farwaniya (+918133066128) Cytotec 200mg tablets Al AHMEDI
Abortion pills Buy Farwaniya (+918133066128) Cytotec 200mg tablets Al AHMEDI
 
ACNE VULGARIS , ALLERGIES, ECZEMA, PEMPHIGUS.pdf
ACNE VULGARIS , ALLERGIES, ECZEMA, PEMPHIGUS.pdfACNE VULGARIS , ALLERGIES, ECZEMA, PEMPHIGUS.pdf
ACNE VULGARIS , ALLERGIES, ECZEMA, PEMPHIGUS.pdf
 
@Safe Abortion pills IN Jeddah(+918133066128) Un_wanted kit Buy Jeddah
@Safe Abortion pills IN Jeddah(+918133066128) Un_wanted kit Buy Jeddah@Safe Abortion pills IN Jeddah(+918133066128) Un_wanted kit Buy Jeddah
@Safe Abortion pills IN Jeddah(+918133066128) Un_wanted kit Buy Jeddah
 
Making change happen: learning from "positive deviancts"
Making change happen: learning from "positive deviancts"Making change happen: learning from "positive deviancts"
Making change happen: learning from "positive deviancts"
 
Post marketing surveillance in Japan, legislation and.pptx
Post marketing surveillance in Japan, legislation and.pptxPost marketing surveillance in Japan, legislation and.pptx
Post marketing surveillance in Japan, legislation and.pptx
 
RESPIRATORY ALKALOSIS & RESPIRATORY ACIDOSIS.pdf
RESPIRATORY ALKALOSIS & RESPIRATORY ACIDOSIS.pdfRESPIRATORY ALKALOSIS & RESPIRATORY ACIDOSIS.pdf
RESPIRATORY ALKALOSIS & RESPIRATORY ACIDOSIS.pdf
 
The Events of Cardiac Cycle - Wigger's Diagram
The Events of Cardiac Cycle - Wigger's DiagramThe Events of Cardiac Cycle - Wigger's Diagram
The Events of Cardiac Cycle - Wigger's Diagram
 
No Advance 931~602~0077 Goa ✂️ Call Girl , Indian Call Girl Goa For Full nig...
No Advance  931~602~0077 Goa ✂️ Call Girl , Indian Call Girl Goa For Full nig...No Advance  931~602~0077 Goa ✂️ Call Girl , Indian Call Girl Goa For Full nig...
No Advance 931~602~0077 Goa ✂️ Call Girl , Indian Call Girl Goa For Full nig...
 
❤️ Chandigarh Call Girls ☎️99158-51334☎️ Escort service in Chandigarh ☎️ Chan...
❤️ Chandigarh Call Girls ☎️99158-51334☎️ Escort service in Chandigarh ☎️ Chan...❤️ Chandigarh Call Girls ☎️99158-51334☎️ Escort service in Chandigarh ☎️ Chan...
❤️ Chandigarh Call Girls ☎️99158-51334☎️ Escort service in Chandigarh ☎️ Chan...
 
2024 PCP #IMPerative Updates in Rheumatology
2024 PCP #IMPerative Updates in Rheumatology2024 PCP #IMPerative Updates in Rheumatology
2024 PCP #IMPerative Updates in Rheumatology
 
CALCIUM - ELECTROLYTE IMBALANCE (HYPERCALCEMIA & HYPOCALCEMIA).pdf
CALCIUM - ELECTROLYTE IMBALANCE (HYPERCALCEMIA & HYPOCALCEMIA).pdfCALCIUM - ELECTROLYTE IMBALANCE (HYPERCALCEMIA & HYPOCALCEMIA).pdf
CALCIUM - ELECTROLYTE IMBALANCE (HYPERCALCEMIA & HYPOCALCEMIA).pdf
 
VIP Just Call 9548273370 Lucknow Top Class Call Girls Number | 8630512678 Esc...
VIP Just Call 9548273370 Lucknow Top Class Call Girls Number | 8630512678 Esc...VIP Just Call 9548273370 Lucknow Top Class Call Girls Number | 8630512678 Esc...
VIP Just Call 9548273370 Lucknow Top Class Call Girls Number | 8630512678 Esc...
 
Obat Penggugur Kandungan Cytotec Dan Gastrul Harga Indomaret
Obat Penggugur Kandungan Cytotec Dan Gastrul Harga IndomaretObat Penggugur Kandungan Cytotec Dan Gastrul Harga Indomaret
Obat Penggugur Kandungan Cytotec Dan Gastrul Harga Indomaret
 
Spauldings classification ppt by Dr C P PRINCE
Spauldings classification ppt by Dr C P PRINCESpauldings classification ppt by Dr C P PRINCE
Spauldings classification ppt by Dr C P PRINCE
 
Goa Call Girl 931~602~0077 Call ✂️ Girl Service Vip Top Model Safe
Goa Call Girl  931~602~0077 Call ✂️ Girl Service Vip Top Model SafeGoa Call Girl  931~602~0077 Call ✂️ Girl Service Vip Top Model Safe
Goa Call Girl 931~602~0077 Call ✂️ Girl Service Vip Top Model Safe
 
TEST BANK For Robbins & Kumar Basic Pathology, 11th Edition by Vinay Kumar, A...
TEST BANK For Robbins & Kumar Basic Pathology, 11th Edition by Vinay Kumar, A...TEST BANK For Robbins & Kumar Basic Pathology, 11th Edition by Vinay Kumar, A...
TEST BANK For Robbins & Kumar Basic Pathology, 11th Edition by Vinay Kumar, A...
 
Coach Dan Quinn Commanders Feather T Shirts
Coach Dan Quinn Commanders Feather T ShirtsCoach Dan Quinn Commanders Feather T Shirts
Coach Dan Quinn Commanders Feather T Shirts
 

Pneumonia diagnosis tool Case Study

  • 1. Client A company developing a healthcare platform hired Elinext to help it build a pneumonia diagnosis tool. Challenge The company has been developing a comprehensive healthcare platform. Treating pneumonia has been one of its focus areas due to COVID, and it wanted to build a pneumonia diagnosis tool for the platform. The tool was destined to analyze lung X-ray images and identify signs of pneumonia using machine learning (ML), an artificial intelligence (AI) technique. The company didn’t have relevant in-house experts, so they reached out for help and found it with Elinext. Solution We began by looking for a neural network that would best analyze lung images and found four candidates: ResNet (50, 101, 152), VGG (16, 19), MobileNet and Inception (V2, V3). After digging deeper into each of them, we chose InceptionV3 developed by Google Research Lab. Once we chose our neural network, we moved on to designing the software architecture and training the algorithm.
  • 2. Architecture The software is based on web technology and can be integrated into other systems like desktop applications and mobile apps. We used publicly available frameworks, libraries and technologies to develop the software. To create a static HTML5 web page, we deployed a web server in a Docker container. On that page, a user can upload a lung image and get feedback. The image is sent for processing through the HTTP protocol. Training Training is the most challenging part in building ML algorithms. Your ability to source enough data, avoid errors and be consistent throughout the process can make or break the algorithm. Manual training is often inconsistent. You may forget which steps you have taken and in which order, or occasionally delete logs. As a result, you won’t be able to accurately repeat a training session. Therefore, we automated the process from A to Z. We needed to train complex models with huge datasets fast. To do that, we rented an Amazon Web Services (AWS) g3s.xlarge instance and used Deep Learning Base AMI (Ubuntu 18.10). The latter is a powerful machine boasting 16GB of RAM, a 4-core CPU and an Nvidia Tesla M60 GPU. It was a perfect fit for the task. Once we have chosen the technology, the training could begin. We built a clean Docker container to isolate the model from outer influences and downloaded a ton of lung images from Kaggle. To be able to work with the images, we subsampled them, narrowing them down to a relevant and consistent selection. The dataset and training environment were ready. The training began. We faced a challenge in overtraining, whereby the model could memorize training images and as a result fail to accurately analyze new images in the future. Our solution was to slightly modify the images’ width, height, graininess and some other parameters. We also launched Tensorboard to monitor training metrics. At the final stages, we exported the model to an H5 file, a format commonly used across industries from healthcare to aerospace, for testing. We tested it manually and automatically, using preset scripts. Accuracy The model we’ve developed has a margin of confidence and uses binary identification. What does this mean? It means if the algorithm identifies 80% of lungs as unaffected, it will say the lungs are healthy. If the figure is below 80%, it will assume the lungs might be affected and require medical attention. How It Works The user opens the web application in their browser, uploads a lung image, sends it to the service and receives feedback. The feedback will show whether the lungs are healthy or if a doctor should take a look at the image. Result The tool we’ve built can help reduce human error in identifying pneumonia. This is particularly useful during the pandemic when doctors are overloaded and might overlook some signs of illness. We can also scale the model up to identify some other diseases. Scaling the model down will help integrate it into other systems, speed things up and allow for the analysis of multiple images simultaneously.