<![CDATA[Examination of passport photos at this time still uses a manual system or requires human labor. In carrying out inspections, of course, this inspection procedure has several limitations, namely human error because officers also have limited manpower to carry out checks. The aim of this research is can recognize someone's face thougharea face covered by objects or accessories. The method used as face detection vizDNN and methods for facial recognition viz FaceNet. For the results of testing the training model FaceNet has produced the best models with 97.48% accuracy to 5191 test image and for image testing is obtained image accuracy of 97% with test images of 412 and for testing in real time u got the result For normal facial conditions as big 90%, facial condition using glasses as big 83%, condition of the face using a maskas big70%, condition of the face using a hatas big 81% and facial condition using all accessories as big 66%.There forethe system has been running as expected and the methodFaceNet has been able to be implemented in real time on a facial recognition system with the accuracy of facial recognition Very good.]]>
Copyrights © 2023
