Machine learning for medical imaging: methodological failures and recommendations for the future
Reads0
Chats0
TLDR
In this article , the authors review roadblocks to developing and assessing methods in computer analysis of medical images and provide recommendations on how to further address these problems in the future, and also discuss on-going efforts to counteract these problems.Abstract:
Research in computer analysis of medical images bears many promises to improve patients' health. However, a number of systematic challenges are slowing down the progress of the field, from limitations of the data, such as biases, to research incentives, such as optimizing for publication. In this paper we review roadblocks to developing and assessing methods. Building our analysis on evidence from the literature and data challenges, we show that at every step, potential biases can creep in. On a positive note, we also discuss on-going efforts to counteract these problems. Finally we provide recommendations on how to further address these problems in the future. read more
Citations
More filters
Journal ArticleDOI
Sources of Irreproducibility in Machine Learning: A Review
TL;DR: The literature is reviewed to provide an overview and a taxonomy in addition to a discussion on the identified sources of irreproducibility and three lines of further inquiry are identified.
Proceedings ArticleDOI
A Consistent and Differentiable Lp Canonical Calibration Error Estimator
TL;DR: A low-bias, trainable calibration error estimator based on Dirichlet kernel density estimates, which asymptotically converges to the true L p calibration error, which enables this novel estimator to tackle the strongest notion of multiclass calibration, called canonical (or distribution) calibration.
Journal ArticleDOI
A Novel Approach to Detect COVID-19: Enhanced Deep Learning Models with Convolutional Neural Networks
Awf A. Ramadhan,Muhammet Baykara +1 more
TL;DR: This study used an approach that uses image cropping methods and a deep learning technique (updated VGG16 model) to classify three public datasets and obtained 97.50% accuracy for multiple classifications and 99.76% for binary classification.
Journal ArticleDOI
Considerations for the Use of Machine Learning Extracted Real-World Data to Support Evidence Generation: A Research-Centric Evaluation Framework
Melissa Estevez,Corey M. Benedum,Chengsheng Jiang,Aaron Cohen,Sharang Phadke,Somnath Sarkar,Selen Bozkurt +6 more
TL;DR: A research-centric evaluation framework for model developers, ML-extracted data users and other RWD stakeholders is proposed, which covers the fundamentals of evaluating RWD produced using ML methods to maximize the use of EHR data for research purposes.
Journal ArticleDOI
Lightweight Multireceptive Field CNN for 12-Lead ECG Signal Classification
Degaga Wolde Feyisa,Taye Girma Debelee,Yehualashet Megersa Ayano,Samuel Rahimeto Kebede,Tariku Fekadu Assore +4 more
TL;DR: A multireceptive field CNN architecture for ECG classification is designed that can improve the performance of ECG signal classification and considers semantic context information within ECG signals with different sizes.
References
More filters
Journal Article
Statistical Comparisons of Classifiers over Multiple Data Sets
TL;DR: A set of simple, yet safe and robust non-parametric tests for statistical comparisons of classifiers is recommended: the Wilcoxon signed ranks test for comparison of two classifiers and the Friedman test with the corresponding post-hoc tests for comparisons of more classifiers over multiple data sets.
Journal ArticleDOI
A survey on deep learning in medical image analysis
Geert Litjens,Thijs Kooi,Babak Ehteshami Bejnordi,Arnaud Arindra Adiyoso Setio,Francesco Ciompi,Mohsen Ghafoorian,Jeroen van der Laak,Bram van Ginneken,Clara I. Sánchez +8 more
TL;DR: This paper reviews the major deep learning concepts pertinent to medical image analysis and summarizes over 300 contributions to the field, most of which appeared in the last year, to survey the use of deep learning for image classification, object detection, segmentation, registration, and other tasks.
Journal ArticleDOI
The file drawer problem and tolerance for null results
TL;DR: Quantitative procedures for computing the tolerance for filed and future null results are reported and illustrated, and the implications are discussed.
Journal ArticleDOI
CONSORT statement: extension to cluster randomised trials
TL;DR: This paper provides updated and extended guidance, based on the 2010 version of the CONSORT statement and the 2008consORT statement for the reporting of abstracts, on how to report the results of cluster randomised controlled trials.
Journal ArticleDOI
Do we need hundreds of classifiers to solve real world classification problems
TL;DR: The random forest is clearly the best family of classifiers (3 out of 5 bests classifiers are RF), followed by SVM (4 classifiers in the top-10), neural networks and boosting ensembles (5 and 3 members in theTop-20, respectively).