PENGENALAN WAJAH PADA GAMBAR DAN VIDEO DENGAN PERBANDINGAN BEBERAPA METODE PENGUKURAN KEMIRIPANWAJAH; FACIAL RECOGNITION ON IMAGE AND VIDEO WITH COMPARISON ON SOME FACE SIMILARITY MEASUREMENTMETHOD

Mulia, Arie A, Faizah

2015 | Skripsi | FMIPA

Face recognition under uncontrolled environmet such as changes of illumination, different facial expression, and aging effect make face recognition becomes a difficult task for computer. Variation of face appearance could reduce the accuracy of a face recognition system because the number and variation of training sample in system is limited. This research present a design of system that learn and acquire people face incrementally from face data under various condition of illumination, facial expression, face direction and other condition under uncontrolled environment. Application of this method aims to improve system's ability to recognize face in various condition under uncontrolled environment. System able to collect face data from live streaming video, video file, or image files that contain face. Accuracy and time are main parameter used to measure performance of face recognition method. Based on this research, some face similarity measurement method such as L1 distance, L2 distance, Local Binary Pattern (LBP), Eigenface, Fisherface, etc. that have been tested using some standard dataset for bencmarking such as AT&T faces, Yale Faces B, etc. L1 distance show fastest searching time and good accuracy on result. Fast processing time by L1 distance achieved by its very simple equation so it minimize processing time to calculate distance between two face vector.

Kata Kunci : face Recognition; L1 distance; L2 distance; Local Binary Pattern; Eigenface; Fisherface.