About this Research Topic
In recent years, Deep learning-based methods have achieved profound success in a large variety of topics related to inverse problems, e.g. regularization methods, theoretical insights in deep networks, and outstanding empirical performances in numerous applications. Such rapid growth is largely transforming and reshaping the field of inverse problems. We are thus highly motivated to establish this Research Topic and aim to deliver a succinct overview of recent progress on the joint force of deep learning and inverse problems.
One major goal of this Research Topic is to provide a broad overview of research on various themes of inverse problems that involve deep neural networks as a crucial ingredient. Some exemplary themes include methodological innovations (e.g. network architectures, training strategies, regularization techniques, etc.), theoretical understanding (e.g. consistency, rates of convergence, regularization properties, etc.), convergence behavior (e.g. stopping criteria, etc.), and real world applications (e.g. biomedical imaging, modeling, etc.). Other themes of research apart from the aforementioned ones are also welcome. This would lead to a collection of papers with rather mixed subjects, and hence we would expect that the submitted papers are written for a general audience.
We encourage theoretical, algorithmic as well as applied papers, and particularly encourage interdisciplinary and collaborative works across branches of sciences.
Recommended topics include but are not limited to the following:
● Image reconstruction and analysis
● Novel network architectures
● Variational, iterative, and cascaded networks
● Unrolled and recurrent networks
● Regularization theory for deep neural networks
● Theory and applications of learned regularizers
● Generative machine learning approaches for inverse problems
● Deep image prior in inverse problems
● Algorithms and theory for bilevel optimization
● Transfer learning in imaging applications
● Applications in medicine and biomedical imaging
● Applications in natural and engineering sciences
One major goal of this Research Topic is to provide a broad overview of research on various themes of inverse problems that involve deep neural networks as a crucial ingredient. Some exemplary themes include methodological innovations (e.g. network architectures, training strategies, regularization techniques, etc.), theoretical understanding (e.g. consistency, rates of convergence, regularization properties, etc.), convergence behavior (e.g. stopping criteria, etc.), and real world applications (e.g. biomedical imaging, modeling, etc.). Other themes of research apart from the aforementioned ones are also welcome. This would lead to a collection of papers with rather mixed subjects, and hence we would expect that the submitted papers are written for a general audience.
We encourage theoretical, algorithmic as well as applied papers, and particularly encourage interdisciplinary and collaborative works across branches of sciences.
Recommended topics include but are not limited to the following:
● Image reconstruction and analysis
● Novel network architectures
● Variational, iterative, and cascaded networks
● Unrolled and recurrent networks
● Regularization theory for deep neural networks
● Theory and applications of learned regularizers
● Generative machine learning approaches for inverse problems
● Deep image prior in inverse problems
● Algorithms and theory for bilevel optimization
● Transfer learning in imaging applications
● Applications in medicine and biomedical imaging
● Applications in natural and engineering sciences
Keywords: Regularization, Deep Learning, Network Design, Generative Networks, Deep Image Prior, Imaging, Bilevel Optimization, Transfer Learning, Tomography, Microscopy.
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.
