Dataset Information

Novel loss functions for ensemble-based medical image classification.

ABSTRACT: Medical images commonly exhibit multiple abnormalities. Predicting them requires multi-class classifiers whose training and desired reliable performance can be affected by a combination of factors, such as, dataset size, data source, distribution, and the loss function used to train deep neural networks. Currently, the cross-entropy loss remains the de-facto loss function for training deep learning classifiers. This loss function, however, asserts equal learning from all classes, leading to a bias toward the majority class. Although the choice of the loss function impacts model performance, to the best of our knowledge, we observed that no literature exists that performs a comprehensive analysis and selection of an appropriate loss function toward the classification task under study. In this work, we benchmark various state-of-the-art loss functions, critically analyze model performance, and propose improved loss functions for a multi-class classification task. We select a pediatric chest X-ray (CXR) dataset that includes images with no abnormality (normal), and those exhibiting manifestations consistent with bacterial and viral pneumonia. We construct prediction-level and model-level ensembles to improve classification performance. Our results show that compared to the individual models and the state-of-the-art literature, the weighted averaging of the predictions for top-3 and top-5 model-level ensembles delivered significantly superior classification performance (p < 0.05) in terms of MCC (0.9068, 95% confidence interval (0.8839, 0.9297)) metric. Finally, we performed localization studies to interpret model behavior and confirm that the individual models and ensembles learned task-specific features and highlighted disease-specific regions of interest. The code is available at https://github.com/sivaramakrishnan-rajaraman/multiloss_ensemble_models.

SUBMITTER: Rajaraman S

PROVIDER: S-EPMC8718001 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundTransfer learning (TL) with convolutional neural networks aims to improve performances on a new task by leveraging the knowledge of similar tasks learned in advance. It has made a major contribution to medical image analysis as it overcomes the data scarcity problem as well as it saves time and hardware resources. However, transfer learning has been arbitrarily configured in the majority of studies. This review paper attempts to provide guidance for selecting a model and TL approaches for the medical image classification task.Methods425 peer-reviewed articles were retrieved from two databases, PubMed and Web of Science, published in English, up until December 31, 2020. Articles were assessed by two independent reviewers, with the aid of a third reviewer in the case of discrepancies. We followed the PRISMA guidelines for the paper selection and 121 studies were regarded as eligible for the scope of this review. We investigated articles focused on selecting backbone models and TL approaches including feature extractor, feature extractor hybrid, fine-tuning and fine-tuning from scratch.ResultsThe majority of studies (n = 57) empirically evaluated multiple models followed by deep models (n = 33) and shallow (n = 24) models. Inception, one of the deep models, was the most employed in literature (n = 26). With respect to the TL, the majority of studies (n = 46) empirically benchmarked multiple approaches to identify the optimal configuration. The rest of the studies applied only a single approach for which feature extractor (n = 38) and fine-tuning from scratch (n = 27) were the two most favored approaches. Only a few studies applied feature extractor hybrid (n = 7) and fine-tuning (n = 3) with pretrained models.ConclusionThe investigated studies demonstrated the efficacy of transfer learning despite the data scarcity. We encourage data scientists and practitioners to use deep models (e.g. ResNet or Inception) as feature extractors, which can save computational costs and time without degrading the predictive power.

Project description:It has now become clear that in silico prediction of ADME (absorption, distribution, metabolism, and elimination) characteristics is an important component of the drug discovery process. Therefore, there has been considerable interest in the development of in silico modeling of ADME prediction in recent years. Despite the advances in this field, there remains challenges when facing the unbalanced and high dimensionality problems simultaneously. In this work, we introduce a novel adaptive ensemble classification framework named as AECF to deal with the above issues. AECF includes four components which are (1) data balancing, (2) generating individual models, (3) combining individual models, and (4) optimizing the ensemble. We considered five sampling methods, seven base modeling techniques, and ten ensemble rules to build a choice pool. The proper route of constructing predictive models was determined automatically according to the imbalance ratio (IR). With the adaptive characteristics of AECF, it can be used to work on the different kinds of ADME data, and the balanced data is a special case in AECF. We evaluated the performance of our approach using five extensive ADME datasets concerning Caco-2 cell permeability (CacoP), human intestinal absorption (HIA), oral bioavailability (OB), and P-glycoprotein (P-gp) binders (substrates/inhibitors, PS/PI). The performance of AECF was evaluated on two independent datasets, and the average AUC values were 0.8574-0.8602, 0.8968-0.9182, 0.7821-0.7981, 0.8139-0.8311, and 0.8874-0.8898 for CacoP, HIA, OB, PS and PI, respectively. Our results show that AECF can provide better performance and generality compared with individual models and two representative ensemble methods bagging and boosting. Furthermore, the degree of complementarity among the AECF ensemble members was investigated for the purpose of elucidating the potential advantages of our framework. We found that AECF can effectively select complementary members to construct predictive models by our auto-adaptive optimization approach, and the additional diversity in both sample and feature space mainly contribute to the complementarity of ensemble members.

Dataset Information

Novel loss functions for ensemble-based medical image classification.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets