OPG-based dental age estimation using a data-technical exploration of deep learning techniques.

Dental age estimation, a cornerstone in forensic age assessment, has been extensively tried and tested, yet manual methods are impeded by tedium and interobserver variability. Automated approaches using deep transfer learning encounter challenges like data scarcity, suboptimal training, and fine-tuning complexities, necessitating robust training methods. This study explores the impact of convolutional neural network hyperparameters, model complexity, training batch size, and sample quantity on age estimation. EfficientNet-B4, DenseNet-201, and MobileNet V3 models underwent cross-validation on a dataset of 3896 orthopantomograms (OPGs) with batch sizes escalating from 10 to 160 in a doubling progression, as well as random subsets of this training dataset. Results demonstrate the EfficientNet-B4 model, trained on the complete dataset with a batch size of 160, as the top performer with a mean absolute error of 0.562 years on the test set, notably surpassing the MAE of 1.01 at a batch size of 10. Increasing batch size consistently improved performance for EfficientNet-B4 and DenseNet-201, whereas MobileNet V3 performance peaked at batch size 40. Similar trends emerged in training with reduced sample sizes, though they were outperformed by the complete models. This underscores the critical role of hyperparameter optimization in adopting deep learning for age estimation from complete OPGs. The findings not only highlight the nuanced interplay of hyperparameters and performance but also underscore the potential for accurate age estimation models through optimization. This study contributes to advancing the application of deep learning in forensic age estimation, emphasizing the significance of tailored training methodologies for optimal outcomes.

Hilbert Vibration Decomposition-based epileptic seizure prediction with neural network.

Epilepsy is one of the most prominent brain disorders in the world, and epileptic patients suffer from sudden seizures that have a substantial negative impact on their lives. A seizure prediction system, therefore, is essential in overcoming the difficulties that epileptic individuals experience. This study designs and demonstrates a non-patient specific seizure prediction system that uses the Hilbert Vibration Decomposition (HVD) method on surface EEG recordings of 10 patients from the CHB-MIT database. EEG signals with 18 channels are decomposed to 7 subcomponents with the HVD in sliding windows. These subcomponents from all channels are then used to calculate features to be fed into an MLP classifier. The classification process is performed for all patients simultaneously and without relaying information concerning patient identity to the classifier. After the classification stage, an alarm algorithm that evaluates the frequency of preictal predictions is developed. The classification sensitivity was, on average, 19.89% across patients. This sensitivity was increased to, on average 89.8% within 120 min and an average false alarm rate of 0.081/h was achieved with a seizure prediction horizon of 4 min across patients after alarm creation.