Detection of hijab syar’i as smart clothes system for moslem people using high performance of parallel computing

Abstract

Moslems are required to remain consistent in wearing syar’i clothing. Sometimes, however, it is very difficult to distinguish between syar’i clothes and the non-syar’i ones, especially in the current millennial era and until the end of time there are many temptations that are not good. Digital image processing in the present day has become one of the areas that can be processed to help various problems associated with images. However, it requires very long processing time and ineffective processing. The CPU alone is not enough. Therefore, it takes processing time allocation that can cut the time to speed up the process. This study utilizes the way to speed up performance analysis of digital images through GPU. The performance analysis of GPU includes CUDA and Yolo because it allows specialists in parallel programming to use GPU resources and can perform the object detection process quickly and accurately by applying an artificial neural network to the image of someone who is wearing syar‘i hijab as smart clothes for Moslem women. Several hardware, software, and file dependency specifications are employed to support the implementation process. The test results achieve the accuracy of 100%, proving that by using the implementation of Yolo under GPU.