Algorithms 2021, 14, 278. https://doi.org/10.3390/a14100278 www.mdpi.com/journal/algorithms Article Ensembling EfficientNets for the Classification and Interpretation of Histopathology Images Athanasios Kallipolitis 1, *, Kyriakos Revelos 2 and Ilias Maglogiannis 1, * 1 Department of Digital Systems, University of Piraeus, 18534 Piraeus, Greece 2 251 Hellenic Air Force and Veterans General Hospital, 11525 Athens, Greece; Kyriakos.revelos@haf.gr * Correspondence: nasskall@unipi.gr (A.K.); imaglo@unipi.gr (I.M.) Abstract: The extended utilization of digitized Whole Slide Images is transforming the workflow of traditional clinical histopathology to the digital era. The ongoing transformation has demonstrated major potentials towards the exploitation of Machine Learning and Deep Learning techniques as assistive tools for specialized medical personnel. While the performance of the implemented algo- rithms is continually boosted by the mass production of generated Whole Slide Images and the de- velopment of state-of the-art deep convolutional architectures, ensemble models provide an addi- tional methodology towards the improvement of the prediction accuracy. Despite the earlier belief related to deep convolutional networks being treated as black boxes, important steps for the inter- pretation of such predictive models have also been proposed recently. However, this trend is not fully unveiled for the ensemble models. The paper investigates the application of an explanation scheme for ensemble classifiers, while providing satisfactory classification results of histopathology breast and colon cancer images in terms of accuracy. The results can be interpreted by the hidden layers’ activation of the included subnetworks and provide more accurate results than single net- work implementations. Keywords: ensemble classifiers; explainability; EfficientNets; digital pathology; whole slide images; guided-grad cam; breast cancer; colon cancer 1. Introduction Machine learning techniques with a dedicated emphasis on deep learning method- ologies have been applied successfully on the field of health informatics as an assistive tool for the relief of workload that specialized medical personnel need to carry [1,2] and for educational purposes [3]. The iterative process of continuously evolving the concerned algorithms has brought to light more effective implementations that exceed the human eye discriminative capability [4–6] and enhance the objectivity criteria by means of visual patterns’ quantification. These improved implementations are, therefore, applied for the reliable and precise prognosis and diagnosis of pathologic cases. The processing of traditional medical imaging material such as MRI’s, X-ray’s, Ultra- sounds, Endoscopy, Thermography, Tomography, Microscopy, and Dermoscopy has been transformed to each digital version providing numerous benefits in a variety of tasks that were earlier performed manually [2,7–13]. The abovementioned tasks fall under the umbrella of well-known computer vision tasks, namely, semantic segmentation [14,15], generation [16], registration [17,18], image classification [15], and object detection [19]. In the last decade, the registered and documented ability of deep convolutional networks to identify visual patterns beyond the human perspective is gaining popularity in the field of digital pathology as well. Driven by the rise of digital scanners that produce whole slide images, the assessment of human tissue in histopathology images can be conducted by means of a virtual microscope. A whole slide image, containing in average 10 GB, can Citation: Kallipolitis, A.; Revelos, K.; Maglogiannis, I. Ensembling EfficientNets for the Classification and Interpretation of Histopathology Images. Algorithms 2021, 14, 278. https://doi.org/10.3390/a14100278 Academic Editors: Panagiotis Pintelas, Ioannis E. Livieris and Frank Werner Received: 22 August 2021 Accepted: 24 September 2021 Published: 26 September 2021 Publisher’s Note: MDPI stays neu- tral with regard to jurisdictional claims in published maps and institu- tional affiliations. Copyright: © 2021 by the authors. Li- censee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and con- ditions of the Creative Commons At- tribution (CC BY) license (http://crea- tivecommons.org/licenses/by/4.0/).