Diagnosis and treatment recommendation engine for skin diseases using AI model VGG19

• Main Authors: Nimitha, N., Aarthi, J., Bhuvaneshwari, M., Brindha, L., Divya, P. L.
• Format: Conference Proceeding
• Language: English
• Subjects: Artificial neural networks; Background noise; Deep learning; Diagnosis; Health services; Image acquisition; Medical imaging; Neural networks; Recommender systems; Skin diseases
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0072492

Skin diseases have a heavy impact on people's life and health. If the skin diseases detected early, the curability is very high and over 92%. In this paper, we analyzed three type skin diseases such as melanoma, impetigo and ringworm. Initially, skin images were pre-processed to get rid of noise and irrelevant background by filtering and transformation. In the Deep learning process, a convolutional neural network is used to analyze visual imagery. VGG-19 (Convolutional neural network) is employed to load the pretrained version of the network trained on quite 1,000,000 images. The Deep learning is capable of make decisions based on the past data and it will use for many diseases identification and treatment. It has changed our way of health monitoring, diagnosis the disease and its level. In this proposal we use the Deep learning to acquire data from dermoscopy image and from that we diagnosis the impact of disease and monitor it on real time process. Based on the data received from the MATLAB we will indicate the types of disease. We determined the accuracy of 96.15% and precession of 94.90% as a result of effective system proposed in the paper.