Deep Learning Understanding Of Skin Disease Management
Akash Saxena, R. Shashikala, Gargibala Satpathy, Eswaramoorthy Manikandan, Asheesh Kumar, Bhasker Pant · 2022 2nd International Conference on Advance Computing and Innovative Technologies in Engineering (ICACITE) · 2022
Deep learning methods are effective in learning characteristics that aid in accurately interpreting complicated patterns. This study presented a deep attempt to learn Mobile Net Setup and Strong Real Limit Memory-based automated approach for identifying skin diseases (LSTM). The Convolution neural network V2 model has proven to be efficient and accurate and it can operate on lightweight computing devices. The suggested model is effective at preserving tasteful information for accurate predictions. To measure the progression of pathological growth, a grey-level founder matrix is utilized. The results were compared to other cutting - edge models such as Exquisite Deep Neural network (FTNN), Convolution Neural Networks (CNN), Very Convolution Neural Networks for Large - scale image Recognizing established by Radial Basis function Group (VGG), and deep neural networks architectural styles that expanded with few changes. The Intermediate representation dataset is employed, and the suggested strategy outperforms existing methods by more than 85%. Its resilience in recognizing the afflicted region significantly faster and with almost twice as few calculations as the typical Convolution neural network model results in minimum computing effort. Additionally, a software program is intended to take immediate and appropriate action. It assists patients and dermatologists in determining the kind of disease from images of the afflicted region during the early stages of skin disease. These findings imply that the suggested method can assist general practitioners in swiftly and effectively diagnosing skin disorders, preventing future complications and morbidity.