The latest developments in the smartphone-based skin cancer diagnosis application allow simple ways for portable melanoma risk assessment and diagnosis for early skin cancer detection. Due to the trade-off problem (time complexity and error rate) on using a smartphone to run a machine learning algorithm for image analysis, most of the skin cancer diagnosis apps execute the image analysis on the server. In this study, we investigate the performance of skin cancer images detection and classification on android devices using the MobileNet v2 deep learning model. We compare the performance of several aspects; object detection and classification method, computer and android based image analysis, image acquisition method, and setting parameter. Skin cancer actinic Keratosis and Melanoma are used to test the performance of the proposed method. Accuracy, sensitivity, specificity, and running time of the testing methods are used for the measurement. Based on the experiment results, the best parameter for the MobileNet v2 model on android using images from the smartphone camera produces 95% accuracy for object detection and 70% accuracy for classification. The performance of the android app for object detection and classification model was feasible for the skin cancer analysis. Android-based image analysis remains within the threshold of computing time that denotes convenience for the user and has the same performance accuracy with the computer for the high-quality images. These findings motivated the development of disease detection processing on android using a smartphone camera, which aims to achieve real-time detection and classification with high accuracy.

Today, the use of mobile devices to support day-to-day communications has become part of public life, including communication by using the Internet through wireless networks. This can broaden the utilization of Mobile Learning (M-Learning). M-Learning can intersect with a web-based E-Learning in terms of data utilization. This article discusses the development of E-Learning capabilities that can be used for M-Learning using E-learning architecture that already exist before. As a result, the architecture of the new E-Learning with the ability of M-Learning is constructed and the prototype based on the architecture is developed.

Mobile Learning ( M -Learning ) is the learning by using mobile devices, regardless of time and place. M-Learning has the constraint that M-Learning users must connect to the internet to be able to get the learning content . With the Android device, that could be implanted a native application inside, allowing the M -Learning content stored on the device . This study examines how the native application can be developed with Personal Extreme Programming method. In addition , this study reinforces previous research that E-Learning can be expanded ability to be M-Learning. As an outcome of this study is a MLearning application that utilizes E-Learning that has existed , in which M -Learning applications that have some capability of E -Learning that capture and store content on mobile devices , so that the content is not required to access the internet connection .

Skin has a great risk to be suffering from disease. Skin disease is easy to see by the other people, which could urge patient to look for health services and medications immediately. However, most people are less conscious about their skin diseases because many new skin disease which not familiar for patient, so that the skin disease can't be handle and become worse. Information technology could solve those problems by capturing data and deliver optimal output using particular processes. This research aims to overcome the problem by developing a web based information system that implements backpropagation neural network. The symptoms of skin diseases are used as inputs and the match skin disease as output. The architecture of backpropagation neural network in this researchhas four input neurons on input layer, a hidden layer with adjustable amount of neurons and an output neuron on output layer. As a result the most optimal recognition value with validity percentage of 100% on data training and 40% on data testing with training time in 6 hours and 10 minutes using 100000 maximum epoch, 0.0001 minimum error, 0.4 learning rate and 20 neurons in hidden layer.

No abstract