Improving the Robustness and Quality of Biomedical CNN Models through Adaptive Hyperparameter Tuning

Abstract

Deep learning is an obvious method for the detection of disease, analyzing medical images and many researchers have looked into it. However, the performance of deep learning algorithms is frequently influenced by hyperparameter selection, the question of which combination of hyperparameters are best emerges. To address this challenge, we proposed a novel algorithm for Adaptive Hyperparameter Tuning (AHT) that automates the selection of optimal hyperparameters for Convolutional Neural Network (CNN) training. All of the optimal hyperparameters for the CNN models were instantaneously selected and allocated using a novel proposed algorithm Adaptive Hyperparameter Tuning (AHT). Using AHT, enables CNN models to be highly autonomous to choose optimal hyperparameters for classifying medical images into various classifications. The CNN model (Deep-Hist) categorizes medical images into basic classes: malignant and benign, with an accuracy of 95.71%. The most dominant CNN models such as ResNet, DenseNet, and MobileNetV2 are all compared to the already proposed CNN model (Deep-Hist). Plausible classification results were obtained using large, publicly available clinical datasets such as BreakHis, BraTS, NIH-Xray and COVID-19 X-ray. Medical practitioners and clinicians can utilize the CNN model to corroborate their first malignant and benign classification assessment. The recommended Adaptive high F1 score and precision, as well as its excellent generalization and accuracy, imply that it might be used to build a pathologist’s aid tool.

SIENNA: Generalizable Lightweight Machine Learning Platform for Brain Tumor Diagnostics

“…The performance of the native model is further optimized by tuning different combinations of hyperparameters for task-specific implementations. HYPERAS and the HYPEROPT library [49] are used to rapidly explore a range of user-defined search spaces, determined through data-driven experimentation and insights from prior literature [59][60][61][62][63]. Details are presented in Methods and the following specific analysis.…”

“…histopathology [62] and with biomedical imaging [63]. SIENNA demonstrates on average accuracy on clinical DICOM MRI data across 3 tasks of 92% (Non-Tumor, SD=5.5%), 91% (GBM, SD = 3.2%), and 93% (MET, SD = 2.6%), with the distribution of accuracies skewed higher to 100% and a lower bound at 75%.…”

“…SIENNA’s ability to identify meaningful patterns and attributes is further enriched by adversarial training, [57,58] using images with subtle parameter distortion to exploit the model’s vulnerabilities and decision boundaries for re-optimization. Finally, we apply hyperparameter tuning [59,60] to SIENNA for robustness in diagnostic output, as has been demonstrated for CNN analysis applied in genetics [59], vision [61], histopathology [62] and with biomedical imaging [63]. SIENNA demonstrates on average accuracy on clinical DICOM MRI data across 3 tasks of 92% (Non-Tumor, SD=5.5%), 91% (GBM, SD = 3.2%), and 93% (MET, SD = 2.6%), with the distribution of accuracies skewed higher to 100% and a lower bound at 75%.…”

Analysis of the Factors Influencing the Predictive Learning Performance Using Federated Learning

“…In our research investigation, we explore the application of the Naïve Bayes algorithm for predicting student grades and effectively categorizing them into four levels: Low, Good, Average, and Drop [43,44]. Naïve Bayes is a well-established probabilistic classification method known for its simplicity, efficiency, and effectiveness, especially in managing high-dimensional datasets.…”

Recent Advancements and Future Prospects in Active Deep Learning for Medical Image Segmentation and Classification

“…DL-based approaches are extensively adopted for pattern recognition, primarily due to their unique feature of being trainable as a complete program [19], [20], [21]. The effectiveness of the deep learning algorithm is predicated on vast and high-quality image datasets to ensure precise learning and training [22], [23]. Utilizing neural networks, deep learning aids in efficient image processing, aiding specialists in diagnosis and enhancing diagnostic accuracy.…”