NeuroNasal: Advanced AI-Driven Self-Supervised Learning Approach for Enhanced Sinonasal Pathology Detection DOI Creative Commons
Nesrine Atitallah, Safa Ben Atitallah, Maha Driss

и другие.

Sensors, Год журнала: 2025, Номер 25(8), С. 2369 - 2369

Опубликована: Апрель 8, 2025

Sinus diseases are inflammations or infections of the sinuses that significantly impact patient quality life. They cause nasal congestion, facial pain, headaches, thick discharge, and a reduced sense smell. However, accurately diagnosing these is challenging due to multiple factors, including inadequate adherence pre-diagnostic protocols. By leveraging latest developments in Artificial Intelligence (AI), there exists substantial opportunity improve precision effectiveness classification diseases. In this study, we present novel AI-based approach for sinonasal pathology detection, using Self-Supervised Learning (SSL) techniques Random Forest (RF) algorithms. We have collected new diagnostic imaging dataset, which major contribution study. The dataset contains 137 CT MRI images meticulously labeled by expert radiologists, with two classes: healthy unhealthy (sinus disease). This useful asset developing evaluating techniques. addition, our proposed employs Deep InfoMax (DIM) model extract meaningful global local features from data self-supervised method. These then used as input an RF classifier, effectively distinguishes between pathological cases. combination both DIM provides efficient feature learning powerful sinus Our preliminary results demonstrate efficiency approach, achieves mean accuracy 92.62%. findings highlight potential improving diagnosis.

Язык: Английский

NeuroNasal: Advanced AI-Driven Self-Supervised Learning Approach for Enhanced Sinonasal Pathology Detection DOI Creative Commons
Nesrine Atitallah, Safa Ben Atitallah, Maha Driss

и другие.

Sensors, Год журнала: 2025, Номер 25(8), С. 2369 - 2369

Опубликована: Апрель 8, 2025

Sinus diseases are inflammations or infections of the sinuses that significantly impact patient quality life. They cause nasal congestion, facial pain, headaches, thick discharge, and a reduced sense smell. However, accurately diagnosing these is challenging due to multiple factors, including inadequate adherence pre-diagnostic protocols. By leveraging latest developments in Artificial Intelligence (AI), there exists substantial opportunity improve precision effectiveness classification diseases. In this study, we present novel AI-based approach for sinonasal pathology detection, using Self-Supervised Learning (SSL) techniques Random Forest (RF) algorithms. We have collected new diagnostic imaging dataset, which major contribution study. The dataset contains 137 CT MRI images meticulously labeled by expert radiologists, with two classes: healthy unhealthy (sinus disease). This useful asset developing evaluating techniques. addition, our proposed employs Deep InfoMax (DIM) model extract meaningful global local features from data self-supervised method. These then used as input an RF classifier, effectively distinguishes between pathological cases. combination both DIM provides efficient feature learning powerful sinus Our preliminary results demonstrate efficiency approach, achieves mean accuracy 92.62%. findings highlight potential improving diagnosis.

Язык: Английский

Процитировано

0