A Decision Support System for Melanoma Diagnosis from Dermoscopic Images

Abstract

Innovative technologies in dermatology allow for the early screening of skin cancer, which results in a reduction in the mortality rate and surgical treatments. The diagnosis of melanoma is complex not only because of the number of different lesions but because of the high similarity amongst skin lesions of different nature; hence, human vision and physician experience still play a major role. The adoption of automatic systems would aid clinical assessment and make the diagnosis reproducible by eliminating inter- and intra-observer variabilities. In our paper, we describe a computer-aided system for the early diagnosis of melanoma in dermoscopic images. A soft pre-processing phase is performed so as to avoid the loss of details both in texture, colors, and contours, and color-based image segmentation is later carried out using k-means. Features linked to both geometric properties and color characteristics are used to analyze skin lesions through a support vector machine classifier. The PH2 public database is used for the assessment of the procedure’s sensitivity, specificity, and accuracy. A statistical approach is carried out to establish the impact of image quality on performance. The obtained results show remarkable achievements, so our computer-aided approach should be suitable as a Decision Support System for melanoma detection.

BFT—Low-Latency Bit-Slice Design of Discrete Fourier Transform

“…In the stated simplified hypothesis, by a simple substitution of Equation (5) in Equation (1) we have…”

“…Moreover, it is a relevant functional block in modern digital communication systems, for applications in orthogonal frequencydivision multiplexing (OFDM) systems [3,4], such as digital worldwide interoperability for microwave access, wireless metropolitan-area networks, and in the areas of next-generation wireless systems, radar signal processing, spectra sensing of cognitive radio, etc. FFT is also used in medical imaging for image filtering, analysis, and image reconstruction [5,6].…”

Toward Early and Objective Hand Osteoarthritis Detection by Using EMG during Grasps

“…The merging of technology and medical science plays an essential role in the prevention, diagnosis and treatment of illnesses and diseases, including patient diagnostic data [17]. Health technology helps clinicians screen abnormalities and contributes to detecting clinical signs [18]. Thus, studying the forearm's muscle signals while performing the most relevant grasps in daily life can lead to the finding of indicators that help detect HOA before the main symptoms appear.…”

Review of methods for early melanoma computer vision detection

“…Классификация изображений на основе алгоритмов глубокого обучения. На этапе предобработки выполняется преобразование изображения к стандартным размерам, нормализация по каждому каналу модели RGB.Сегментация изображения может выполняться на основе цветовой оптимизации каналов изображения[28,29], методов фиксации уровня[30], методы формирования однородных областей[31], применения глубоких нейронных сетей[32][33][34].При обнаружении меланомы используются нейронные сети на основе архитектуры EfficientNet[35].…”