Nabilah Hanun, Meochammad Sarosa, Rosa Andrie Asmara
Human detection is one of technology that could be implemented in many ways such as security, crime prevention, accidents, absenteeism, victims of natural disaster discovery, and many more. Humans have different shapes and sizes influenced by genetics and life patterns, so human detection technology is considered attractive to be implemented with various existing methods. In its development, human detection has been carried out in several different methods ranging from traditional to the most effective. This study uses the Faster R-CNN algorithm method with Inception ResNet v2 network architecture. Based on the tests that have been carried out performances that the improved network can effectively improve the efficiency of network operations, after testing 7 times ranging from 3000 steps to 13 steps, the accuracy of recognition of human body cut objects reached the highest 81.30% at 9000 steps Testing with a loss of 0.13671. In this way, it shows satisfactory results in agreeing with pieces of the human body and can be developed with better results. © 2023 IEEE.
Jurusan Teknik Elektro Politeknik Negeri Malang, Malang, Indonesia; Jurusan Teknologi Informasi Politeknik Negeri Malang, Malang, Indonesia