Interpretable deep learning (IDL) applications in instantaneous smart healthcare services

Introduction:

The emergence of cutting-edge technologies is reshaping our approach to everyday life. Deep learning (DL), the Internet of Things (IoT) and cloud computing, have had an impact on our daily lives, from the workplace to the community. In particular, the healthcare industry has been significantly influenced, allowing practitioners to rely on such innovative ideas not only in acquiring precise diagnoses, but also in selecting, administering and assessing interpretable solutions of healthcare services over time. Furthermore, the use of connected sensing devices in conjunction with data sciences creates a slew of new possibilities in the framework of the healthcare system.

DL has been utilized with positive outcomes in recent times for a variety of medical computer vision applications, such as cancer detection, tumor segmentation, malignant categorization, vascular profiling and tumor prognosis. Despite their potential accuracy, DL models remain perceived by physicians and radiologists as complex and opaque. These systems are also susceptible to cyberattacks due to their input sequence dependence. Thus, it is crucial to determine its trustworthiness when it comes to the trained model's capability for accurate data analysis.

Consequently, research in this area focuses on obtaining insight into these models; on how they can fit the data better along with making them more explainable. Nonetheless, there exist questions on what inputs should be given to DL algorithms and how they should operate so that adroit and secure assessments can be accomplished.

To that end, professionals worldwide are interested in learning about the most recent research advances in interpretable deep learning (IDL). This critical research area will examine both emerging difficulties and solutions to future medical research. Multimodal imaging, computerized tomography, magnetic resonance imaging, ultrasound, single photon emission computed tomography, positron emission tomography, optical microscopy, tomography, together with other noninvasive radiological imaging, can be used to collect medical data.

Topics covered:

This special issue aims to highlight transparent medical/healthcare services related to IDL. Potential topics of this research topic should include, but not be limited to:

• IDL applications for medical image processing and visualization
• Healthcare data analysis using IDL
• Intrepretable distributed and transfer learning applications in medical imaging
• Securing against adversarial cyberattacks in healthcare computer vision algorithms
• Innovations in information retrieval, data analytics, and DL in smart healthcare services using IDL
• Advanced methodologies for effective diagnosis using IDL
• IDL models for multimodal heathcare data fusion
• Data warehouse strategies and privacy applied to the mass storage of medical image analysis using IDL
• Big data analytics for real-time healthcare services using IDL
• Re-inforcement learning and federated learning applications in smart healthcare using IDL
• Security and privacy in medical services

Important deadlines:

• Submission deadline: 30^th December 2023

Submission instructions:

Please read the Guide for Authors before submitting. All articles should be submitted online; please select IDL-healthcare upon submission.

Guest editors:

• Dr. Farhan Ullah, Associate Professor, Northwestern Polytechnical University, China. Email: farhan@nwpu.edu.cn
• Dr. Chunwei Tian, Associate Professor, Northwestern Polytechnical University, China. Email: chunweitian@nwpu.edu.cn
• Dr. Syed Aziz Shah, Associate Professor, Coventry University, UK.  Email: syed.shah@coventry.ac.uk
• Dr. Sohail Jabbar, Associate Professor, Imam Mohammad Ibn Saud Islamic University, Saudi Arabia. Email: sjjabar@imamu.edu.sa

For further enquiries, please send emails to dsm@xjtu.edu.cn, and the editing team will answer promptly.