Abstract
The integration of functional imaging modality provided by Positron Emission Tomography (PET) and associated anatomical imaging modality provided by Computed Tomography (CT) has become an essential procedure both in the evaluation of different types of malignancy and in radiotherapy planning. The alignment of these two exams is thus of great importance. In this research work, three registration approaches (1) intensity-based registration, (2) rigid translation followed by intensity-based registration and (3) coarse registration followed by fine-tuning were evaluated and compared. To characterize the performance of these methods, 161 real volume scans from patients involved in Hodgkin Lymphoma staging were used: CT volumes used for radiotherapy planning were registered with PET volumes before any treatment. Registration results achieved 78%, 60%, and 91% of accuracy for methods (1), (2) and (3), respectively. Registration methods validation was extended to a corresponding landmarks points distance calculation. Methods (1), (2) and (3) achieved a median improvement registration rate of 66% mm, 51% mm and 70% mm, respectively. The accuracy of the proposed methods was further confirmed by extending our experiments to other multimodal datasets and in a monomodal dataset with different acquisition conditions. © 2018, Springer Nature Switzerland AG.

Abstract

Abstract

Abstract
Machine learning algorithms have evolved by exchanging simplicity and interpretability for accuracy, which prevents their adoption in critical tasks such as healthcare. Progress can be made by improving interpretability of complex models while preserving performance. This work introduces an extension of interpretable mimic learning which teaches in-terpretable models to mimic predictions of complex deep neural networks, not only on binary problems but also in ordinal settings. The results show that the mimic models have comparative performance to Deep Neural Network models, with the advantage of being interpretable.

Abstract

Abstract
Although cross-validation is a standard procedure for performance evaluation, its joint application with oversampling remains an open question for researchers farther from the imbalanced data topic. A frequent experimental flaw is the application of oversampling algorithms to the entire dataset, resulting in biased models and overly-optimistic estimates. We emphasize and distinguish overoptimism from overfitting, showing that the former is associated with the cross-validation procedure, while the latter is influenced by the chosen oversampling algorithm. Furthermore, we perform a thorough empirical comparison of well-established oversampling algorithms, supported by a data complexity analysis. The best oversampling techniques seem to possess three key characteristics: use of cleaning procedures, cluster-based example synthetization and adaptive weighting of minority examples, where Synthetic Minority Oversampling Technique coupled with Tomek Links and Majority Weighted Minority Oversampling Technique stand out, being capable of increasing the discriminative power of data.