Deep-Learning-Based Scalp Image Analysis Using Limited Data

Abstract

The World Health Organization and Korea National Health Insurance assert that the number of alopecia patients is increasing every year, and approximately 70 percent of adults suffer from scalp problems. Although alopecia is a genetic problem, it is difficult to diagnose at an early stage. Although deep-learning-based approaches have been effective for medical image analyses, it is challenging to generate deep learning models for alopecia detection and analysis because creating an alopecia image dataset is challenging. In this paper, we present an approach for generating a model specialized for alopecia analysis that achieves high accuracy by applying data preprocessing, data augmentation, and an ensemble of deep learning models that have been effective for medical image analyses. We use an alopecia image dataset containing 526 good, 13,156 mild, 3742 moderate, and 825 severe alopecia images. The dataset was further augmented by applying normalization, geometry-based augmentation (rotate, vertical flip, horizontal flip, crop, and affine transformation), and PCA augmentation. We compare the performance of a single deep learning model using ResNet, ResNeXt, DenseNet, XceptionNet, and ensembles of these models. The best result was achieved when DenseNet, XceptionNet, and ResNet were combined to achieve an accuracy of 95.75 and an F1 score of 87.05.

Mobile Application Based on Convolutional Neural Networks for Pterygium Detection in Anterior Segment Eye Images at Ophthalmological Medical Centers

“…Similarly, it had better result in the evolution of loss and accuracy during the training process compared to the other models. These results demonstrate that ResNext50 was also superior to the other evaluated architectures in the field of classification of ophthalmic diseases, which aligns with the findings observed in the literature review [8], [23], [24], [25].…”

“…Additionally, a relatively new architecture called ResNext50, a variant of ResNet50, was found with interesting results. This architecture has been applied outside the ocular domain, as seen in [23], [24], [25], where it achieved better results than other cited models. In the study [23], the model achieved a 99% accuracy and a maximum of 100% precision and recall in detecting harmful algae, surpassing MobileNet-V2, VGG16, and AlexNet.…”

“…In the study [23], the model achieved a 99% accuracy and a maximum of 100% precision and recall in detecting harmful algae, surpassing MobileNet-V2, VGG16, and AlexNet. In [25], ResNext101 was combined with DenseNet169 and XceptionNet41, resulting in an F1 score of 87.74% and an accuracy of 95.75% in detecting alopecia. In [24], on the other hand, it achieved an accuracy of 99.32%, specificity of 99.01%, and recall of 100% in detecting COVID-19, surpassing AlexNet, GoogleNet, ResNet-50, Se-ResNet-50, DenseNet121, Inception V4, and Inception ResNet V2.…”

Development of the information society in Poland on the example of households

“…Współcześnie, technologie takie jak chmura obliczeniowa, big data, Internet rzeczy (IoT) i sztuczna inteligencja (AI) są na czołówce innowacji w ICT. Chmura umożliwia przechowywanie ogromnych ilości danych i dostęp do mocy obliczeniowej na żądanie, IoT łączy fizyczny świat z cyfrowym poprzez inteligentne urządzenia, a AI przyczynia się do tworzenia inteligentnych systemów zdolnych do uczenia się i podejmowania decyzji (Kim, Gil, Kim, Kim, 2023).…”

“…Today, technologies such as cloud computing, big data, the Internet of Things (IoT) and artificial intelligence (AI) are at the forefront of ICT innovation. Clouds make it possible to store of vast amounts of data and provide access to computing power on demand, IoT connects the physical world to the digital one through smart devices, and AI contributes to the creation of intelligent systems capable of learning and decision-making (Kim, Gil, Kim, Kim, 2023).…”