Image Processing Model Based E-Learning for Students Authentication

E-Learning in the Indonesian education community has been growing positively as an electronic information technology application through an internet network designed for the benefit of learning. But it still raises some obstacles in the implementation, as it relates to equity and access. In another aspect, e-learning also contains major weaknesses, namely the decrease in the frequency of direct contact between learners and between students with lecturers and other learning resources, so that learning does not experience completeness in all aspects of cognitive and non-cognitive. The weakness is also accompanied by the suspicion of the institution to the honesty of learners in carrying out the learning process. This study aims at building an e-learning model that can bring the intensity and capacity of learners actually through the virtual world through self-assessment based on image processing. The proposed steps are break into several parts: create a dataset of faces that can be used to evaluate given algorithm. Subsequently the enhancement using histogram equalization allows a strong enhancement on facial features. In addition, then the feature descriptor are selected using Viola and Jones [1]. Afterwards, those features will be saved to database. The next step is to build the system than allows the online learners are detected through our image processing approach then all the interaction will be verified using our proposed infrastructure. The results of the study: (a) Image Processing Based E-Learning model was built with a system capable of running in existing infrastructure so far, and (b) Image Processing Based E-Learning model proved valid and reliable both substantially, system and feasibility significant. The results of this study have implications that can be tested on a larger scale that is for some courses and in a particular department. © 2017 IEEE.