AI based Detection of Acute Respiratory Distress Syndrome (AI-DARDS)

Purpose and goal: ** Denna text är maskinöversatt **

Dielectric profiling: We developed the method "A Fast Domain-Inspired Unsupervised Method to Compute COVID-19 Severity Scores from Lung CT" for profiling.

Development by AI-DARDS-system: Our team created AI-DARDS, a microwave system to diagnose and predict disease severity using machine learning. Phantom model: A fluid-based model with pig lung was developed to validate AI-DARDS and showed high accuracy for ARDS. Data collection: Data from experiments were collected, with future plans for testing in patients.

Expected results and effects: ** Denna text är maskinöversatt **

Acute respiratory distress syndrome (ARDS) has become increasingly acute after COVID-19 due to its rapid development and life-threatening complications. Our research developed a portable microwave system to diagnose and assess ARDS. The system categorizes ARDS into four levels by analyzing the lung´s dielectric changes. En XGBoost-klassificerare improved accuracy. The system offers a non-invasive, portable solution for continuous ARDS monitoring.

Approach and implementation: ** Denna text är maskinöversatt **

Sensor development: We designed a low-profile antenna with directional radiation and broadband frequency for applications such as indoor communication. The design uses a ring-shaped reflector (RBR) to combine low profile and broadband response. An arc-shaped antenna, surrounded by a metal ring, was optimized for frequencies 1.5-3.13 GHz, including 2.45 GHz ISM-bandet.

Phanton development: A liquid phantom was created to mimic the dielectric properties of the chest and verified with DAK 3.5 and VNA.