Giter Club home page Giter Club logo

jufo2023's Introduction

Alzheimer-Erkennung durch Künstliche Intelligenz - JuFo2023

Kurzfassung


German
Alzheimer ist eine zunehmend auftretende neurodegenerative Erkrankung. In ihrem Verlauf nimmt die Fähigkeit des Gehirns Informationen zu speichern ab, die Sprachfähigkeit leidet. Durch eine frühe Erkennung kann das Fortschreiten der Erkrankung entscheidend verlangsamt werden, sodass für Betroffene und Angehörige eine bessere Situation geschaffen wird.
In meinem Projekt möchte ich mittels KI-basierter Ansätze die Diagnostik von Alzheimer in allen Stadien der Erkrankung verbessern. Hierzu verwendete ich MRT-Aufnahmen der Gehirne verschiedener Personen, mithilfe derer ich unterschiedliche Modelle trainierte und vergleichen konnte. Als KI-Architekturen wird ein CNN-Transformer verwendet, der mit unterschiedlich großen ResNets und EfficientNets verglichen wird. Der CNN-Transformer performt bei der Klassifizierung am besten, da er auf Lokalität und Globalität spezialisiert ist.
Durch Methoden der Erklärbarkeit kann zudem die Entscheidung des Modells interpretiert und nachvollzogen werden.

English
Alzheimer's is an increasingly common neurodegenerative disease. In its course, the brain's ability to store information decreases and the ability to speak suffers. Early detection can decisively slow down the progression of the disease, creating a better situation for those affected and their relatives.
Here in this project, I want to use AI-based approaches to improve the diagnosis of Alzheimer's in all stages of the disease. To do this, I used MRI images of the brains of different people, with the help of which I was able to train and compare different models. The AI architectures used are a CNN-Transformer, which is compared with ResNets and EfficientNets of different sizes. The CNN transformer performs best in the classification because it is specialised in locality and globality. In addition, the decision of the model can be interpreted and understood through methods of Explainability. The model and the Explainability methods are combined in an app to help doctors diagnose Alzheimer's patients.
I am also trying to find out more about the gene mechanisms in Alzheimer's disease by clustering cells and genes using Autoencoders and Graph Neural Networks to show the difference in cells from healthy patients to cells from patients with Alzheimer's disease.


Links to the Reports:


Report zu der Hyperparameter Optimierung des CNN-Transformer

Weights&Biases-Reports


Verwenden der App



1. Python in einem Terminal öffnen

2. Installieren der Requirements:

pip install -r requirements.txt

3. Starten der App:

python app.py

4. Aufrufen des folgenden Links in einem Browser.

http://127.0.0.1:8050


Funktionen



Labeled MRT-Regions

MRT-Regions

Overlay interactive Alzheimer-Scan

MRT-Regions

Two labeled Atlases with different depth

MRT-Regions

View MRI-Slices by changing the axial coordinate

MRT-Regions


Upgraded AutoClass


Normal Clustering

NormalClustering

Upgraded AutoClass Clustering

UpgradedAutoClassClustering

Upgraded Disease Clustering

UpgradedAutoClassClustering

Distribution differences with activation functions

scRNA-seq normal ReLU 30 ReLU 10 ReLU 100 Softplus
Non-Zero-Count 5.575.264 8.133.753 1.577.998 7.585.000 36.027.458
Minimum 0.11 2e-07 0.0003 0.006 1e-45
Maximum 24.64 2.74 3.14 2.78 2.58

DistributionDifferences

jufo2023's People

Contributors

niklas1225 avatar

Stargazers

Maximilian Nasert avatar  avatar AperGra avatar Clicks avatar Simon Ma avatar  avatar Paul Brachmann avatar Robert Schwarzenberg avatar

Watchers

 avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.