article · Machine Learning and Knowledge Extraction
This review evaluates convolutional neural network methods developed to automate the manual, expensive and error-prone analysis of medical images. It surveys architectural improvements, frameworks, activation functions, ensemble methods, hyperparameter optimisation, performance metrics, datasets and data preprocessing approaches essential for designing reliable models. In addition, machine learning algorithms were applied to statistically model the existing literature, identifying latent topics, methodological gaps and emerging trends. The analysis highlights a distinct transition from traditional convolutional neural network architectures to hybrid convolutional-transformer models. A notable increase in medical imaging publications was largely spurred by practitioners addressing diagnostic challenges during the COVID-19 pandemic. However, critical gaps remain regarding domain-specific architectural design, post hoc model explainability, and the adaptation of these diagnostic tools for low-resource medical imaging settings.
Interpreting medical images by hand is demanding, expensive and vulnerable to human error. Tracking the development of automated computer vision models helps researchers understand which tools are maturing and where major limitations remain, especially in terms of diagnostic reliability, model interpretability, and practical deployment in resource-constrained healthcare facilities.
The reviewed approaches support automated diagnostic tools that could assist clinical practitioners and radiomics developers in identifying pathological conditions. Because the review highlights persistent gaps in model explainability, optimization, and low-resource viability, the overall field represents early-stage to transitional research rather than fully mature, near-market software ready for immediate clinical integration.
AI-generated from the published abstract. Always read the original work before citing.
In this review, we compiled convolutional neural network (CNN) methods which have the potential to automate the manual, costly and error-prone processing of medical images. We attempted to provide a thorough survey of improved architectures, popular frameworks, activation functions, ensemble techniques, hyperparameter optimizations, performance metrics, relevant datasets and data preprocessing strategies that can be used to design robust CNN models. We also used machine learning algorithms for the statistical modeling of the current literature to uncover latent topics, method gaps, prevalent themes and potential future advancements. The statistical modeling results indicate a temporal shift in favor of improved CNN designs, such as a shift from the use of a CNN architecture to a CNN-transformer hybrid. The insights from statistical modeling point that the surge of CNN practitioners into the medical imaging field, partly driven by the COVID-19 challenge, catalyzed the use of CNN methods for detecting and diagnosing pathological conditions. This phenomenon likely contributed to the sharp increase in the number of publications on the use of CNNs for medical imaging, both during and after the pandemic. Overall, the existing literature has certain gaps in scope with respect to the design and optimization of CNN architectures and methods specifically for medical imaging. Additionally, there is a lack of post hoc explainability of CNN models and slow progress in adopting CNNs for low-resource medical imaging. This review ends with a list of open research questions that have been identified through statistical modeling and recommendations that can potentially help set up more robust, improved and reproducible CNN experiments for medical imaging.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.3390/make6010033
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
