Editorial: Artificial intelligence in point of care diagnostics

Massimo Martinelli, Davide Moroni, Ales Procházka, Marija Strojnik

Abstract: interdisciplinary impact of AI is closely related with the general area of digital signal processing, forming an integrating platform for different applications and unifying their background based on computational intelligence and Machine Learning (ML). This approach follows ideas of Leibnitz presented in history, trying to interconnect researchers of different narrow areas who lost their ability to communicate Prochazka et al. (2021).Indeed, AI and ML can lead to methods for integrating, analyzing and understanding multimedia data from a plethora of different devices. In addition, multivariate methods can correlate the current patient status with the previous history, adapting the findings to his personal history, in line with a more personalized and adaptive approach to care and favoring a more accurate prediction of future status.To this end, there is the need to explore different research directions in AI and PoC Diagnostics. From one side, AI paradigms can be embedded into PoC testing devices, extending their capabilities and making possible analyses otherwise not viable, e.g. those including image analysis. This can lead to a convergence of pervasive computing and PoC Diagnostics. Similarly, networks of local devices can be devised taking advantage of distributed AI: wearable sensors and portable devices can communicate in an ecosystem, and their data can be cumulatively and coherently processed. Finally, AI can be decentralized, also considering a cloud-based approach, extending the capabilities of PoC Diagnostics all over the computational continuum.For instance, with a timely decentralized survey, PoC may allow the detection of anomalies that, once integrated with previously collected data and anamnesis, with the further purpose of a quality check to use reliable data, can be classified by AI methods. Immediately, the system can then alert the user and his caregivers. Moreover, specific assistance networks can guarantee control and rescue over the territory. Even if the cost of PoC devices is high, it reduces indirect costs and saves lives.On the basis of such consideration, it has been our aim to collect in a Research Topic multidisciplinary contributions to "Artificial Intelligence in Point of Care Diagnostics" and, eventually, after a careful revision, five papers were included and published. safe options to diagnose disease, in particular pneumonia. A challenge to the currently existing diagnoses 31 of the pneumonia models has been the feature extraction from the clinical pneumonia X-ray dataset. Four