Deep learning based melanoma diagnosis using dermoscopic images

Abstract

The most common malignancies in the world are skin cancers, with melanomas being the most lethal. The emergence of Convolutional Neural Networks (CNNs) has provided a highly compelling method for medical diagnosis. This research therefore conducts transfer learning with grid search based hyper-parameter fine-tuning using six state-ofthe-art CNN models for the classification of benign nevus and malignant melanomas, with the models then being exported, implemented, and tested on a proof-of-concept Android application. Evaluated using Dermofit Image Library and PH2 skin lesion data sets, the empirical results indicate that the ResNeXt50 model achieves the highest accuracy rate with fast execution time, and a relatively small model size. It compares favourably with other related methods for melanoma diagnosis reported in the literature.

Potential for Chemistry in Multidisciplinary, Interdisciplinary, and Transdisciplinary Teaching Activities in Higher Education

“…ML/AI approaches such as deep learning correlation analysis − can be used to accelerate identification of biological pathways and targets in disease biology, to find potential drug molecules, to understand the biological effects of the compound, and to help design clinical trials to ensure the best outcomes are achieved. Clearly, a multitude of other opportunities exist for inclusion of computational studies in a variety of disciplinary contexts (e.g., agrochemical/drug delivery/design, , catalysis, materials chemistry, medical imaging and analysis, , etc.). Computational chemistry can help engage students learning about important concepts such as kinetics and thermodynamics, molecular descriptors (i.e., constitutional, electronic, physicochemical, topological), stereochemistry, and 3D structures in an interactive virtual learning environment and, moreover, provides various platforms to study intramolecular and intermolecular interactions that can be used for research and development in a safe and cost-effective fashion (e.g., informing the selection of candidates to bind to biological receptors or pharmaceutical carriers for drug delivery, while minimizing resource utilization and exposure to chemicals) .…”

A Deep Ensemble Neural Network with Attention Mechanisms for Lung Abnormality Classification Using Audio Inputs

“…Deep learning has emerged as one of the most popular techniques for video and signal processing tasks, owing to its recent breakthrough and advancement [ 1 ]. A significant number of deep architectures have been proposed for medical diagnosis with respect to diabetic retinopathy screening, brain tumor, and leukemia diagnosis.…”

“…Recurrent Neural Networks (RNNs) have also shown effectiveness for temporal feature extraction with respect to language generation and audio classification. As two popular types of RNNs, Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) have been widely adopted in signal classification and time series forecasting [ 1 , 4 ]. These networks are similar in functionality, with the primary difference being that the GRU combines the “forget” and “input” gates into an “update” gate, as well as having a “reset” gate, instead of an “output” gate as in LSTM.…”

Deep Recurrent Neural Networks with Attention Mechanisms for Respiratory Anomaly Classification

“…Deep learning methods are now one of the most prominent methods in computing today with respect to tasks such as audio, video, and image classification [1]. In the field of medical imaging, a large number of deep learning methods have been demonstrated in a variety of studies.…”