A trustworthy neural architecture search framework for pneumonia image classification utilizing blockchain technology

A chest X-ray radiography is still the global standard for diagnosing pneumonia. Despite several studies, doctors still have trouble correctly diagnosing and classifying pneumonia. Neural architecture search (NAS) has the potential to enhance diagnostic efficiency and accuracy. However, NAS methods fail to account for the security of data sources, and the result of model prediction cannot be communicated safely and consistently. To tackle these issues, we propose a trustworthy NAS method for pneumonia image classification using blockchain technology, which provides secure, reliable, and high-performance model automatic search and efficient data prediction capabilities. By synergistically combining NAS with blockchain technology, we enhance the transparency and interpretability of NAS-driven image classification processes, thereby safeguarding the confidentiality and integrity of sensitive medical information. Moreover, our approach automates the model construction process for pneumonia image classification, markedly reducing the reliance on manual intervention. Experimental results demonstrate that our method achieves comparable performance to state-of-the-art methods for pneumonia image classification while ensuring security. This provides a new solution for promoting medical aided diagnosis.