In recent years, the face recognition field has attained high performance in the identification of a person. Facial recognition algorithms should be able to perform even in the case of similar faces such as Look-alikes or identical twins. Twin identification becomes an important task in face recognition as twins are involved in pursuing criminal activities. The proposed framework focuses on the recognition of individual faces, and Look-alikes by finding distinctiveness of different facial features in the face by using multi-parametric anthropometry measurements. Fusion score is generated by considering the fusion of facial ratios such as image ratio and golden ratio using which the face has been identified. This study aims to explore the relationship of facial features concerning image ratio and golden ratio for facial image comparison. The main objective is to ultimately develop a face recognition system in giving a correct matching decision for recognition of Individual faces of the same person, different person, Look-alikes, and Monozygotic Twins (Identical Twins) which exactly mimics the forensic examiner’s way of facial image comparison which can be used in forensic sciences and be able to present it as statistical evidence.
"Facial image analysis using Ratio-Fusion-Score based approach to support Forensic Investigation,"
Manipal Journal of Science and Technology: Vol. 5:
2, Article 1.
Available at: https://impressions.manipal.edu/mjst/vol5/iss2/1