Training use case for the FIWARE Global Summit 2024 in Naples, Italy.
[
This is a use case example for health domain. The idea of this Proof of Concept (PoC) is to show how a medical image can be integrated into a Powered by FIWARE solution along with a Machine Learning (ML) model in order to perform predictions on observed images.
Any NGSI version could be used.
This Application is for Demonstration Purposes Only
This application is a Proof of Concept (PoC) designed to demonstrate the underlying technology, including Machine Learning, Deep Learning, Keras, Python, FIWARE, JSON, and other technical components.
Important: This application is not intended to be used as a medical tool or to provide medical advice.
- DO NOT use this application to make decisions about your or others health.
- DO NOT rely on this application for the diagnosis or treatment of any medical condition.
For any concerns regarding your health, skin moles, or any other medical issues, please consult a qualified healthcare professional.
Your health and well-being should always be managed by a licensed medical professional. This application is purely for demonstrating the potential of the technology and should not be used as a substitute for professional medical advice, diagnosis, or treatment.
To download this tutorial, simply clone this repository:
gh repo clone dncampo/FGS2024_ML-trainingor just
git clone https://github.com/dncampo/FGS2024_ML-training.gitIf you want to create the CNN Models to perform inference, you should checkout the proper repository and put the models you are going to use under ml-models directory:
git clone https://github.com/dncampo/cheetah.gitOr simply, just download the models I'm using and put them under the mentioned directory.
[Binary model]
[Multiclass model]
To start the containers:
./services [start | stop | build]Will start or stop services - docker containers - respectively. With create param you should be able to download and / or create the containers without starting them.
source ~/venv/bin/activate
To start the front end:
source ~/venv/bin/activate;
streamlit run app.pyWill activate the proper environent and start the Streamlit front end app.
[ngsiLdDefinitionLink]: https://www.etsi.org/deliver/etsi_gs/CIM/001_099/009/01.08.01_60/gs_cim009v010801p.pdfIf anybody is missing, please let me know.