Detection of Measles Using Raspberry Pi 3 with Support Vector Machine

The device is first to go under training using the Support Vector Machine with the sample images to be converted to HSV color space image conversion to differentiate measles from different skin diseases. In detecting measles, input images are used in Support Vector Machine based feature detection algorithm that is developed from the feature map of the training process and to classify the severity based on area and color. The average ratio of measles rash pixels to skin pixels are: (1) light = 2.82% - 20.41%; (2) moderate = 20.42% - 35.03%; (3) severe = 35.04 % and above.