In response to the urgent need for rapid and accurate COVID-19 diagnosis, a two-stage framework is presented to improve the detection of COVID-19 from CT scans. By utilizing annotated data from one domain and non-annotated data from another, the model overcomes challenges of data scarcity and variability common in health crises. The innovative use of pseudo labels enables iterative refinement of the learning process, enhancing accuracy and adaptability across different medical centers. Experimental results on COV19-CT-DB database demonstrate high diagnostic precision, aiding efficient patient management and reducing strain on healthcare systems. Deep-learning methods have been explored for COVID-19 detection, but most require large amounts of annotated data challenging to acquire in real-world scenarios. Employing pseudo labels for domain adaptation addresses this challenge by refining learning processes with unlabeled data, improving predictive accuracy and robustness. The approach bridges gaps between domains, enhancing generalizability across diverse clinical settings.
To Another Language
from source content
arxiv.org
Principais Insights Extraídos De
by Runtian Yuan... às arxiv.org 03-19-2024
https://arxiv.org/pdf/2403.11498.pdfPerguntas Mais Profundas