Back to Search
Start Over
Likelihood Ratios for Deep Neural Networks in Face Comparison
- Source :
- Journal of Forensic Sciences, 65(4), 1169-1183. Wiley-Blackwell, Journal of Forensic Sciences
- Publication Year :
- 2020
- Publisher :
- Wiley-Blackwell, 2020.
-
Abstract
- In this study, we aim to compare the performance of systems and forensic facial comparison experts in terms of likelihood ratio computation to assess the potential of the machine to support the human expert in the courtroom. In forensics, transparency in the methods is essential. Consequently, state‐of‐the‐art free software was preferred over commercial software. Three different open‐source automated systems chosen for their availability and clarity were as follows: OpenFace, SeetaFace, and FaceNet; all three based on convolutional neural networks that return a distance (OpenFace, FaceNet) or similarity (SeetaFace). The returned distance or similarity is converted to a likelihood ratio using three different distribution fits: parametric fit Weibull distribution, nonparametric fit kernel density estimation, and isotonic regression with pool adjacent violators algorithm. The results show that with low‐quality frontal images, automated systems have better performance to detect nonmatches than investigators: 100% of precision and specificity in confusion matrix against 89% and 86% obtained by investigators, but with good quality images forensic experts have better results. The rank correlation between investigators and software is around 80%. We conclude that the software can assist in reporting officers as it can do faster and more reliable comparisons with full‐frontal images, which can help the forensic expert in casework.
- Subjects :
- Paper
Computer science
Automated Facial Recognition
Kernel density estimation
ENFSI proficiency test
Sensitivity and Specificity
01 natural sciences
Convolutional neural network
Pathology and Forensic Medicine
03 medical and health sciences
0302 clinical medicine
Software
Genetics
face verification likelihood ratio
Humans
030216 legal & forensic medicine
Parametric statistics
Likelihood Functions
Commercial software
Models, Statistical
business.industry
Deep learning
Forensic Sciences
010401 analytical chemistry
Nonparametric statistics
deep learning
Confusion matrix
Pattern recognition
0104 chemical sciences
digital forensic science
Papers
Digital & Multimedia Sciences
Neural Networks, Computer
Artificial intelligence
business
face recognition
Subjects
Details
- Language :
- English
- ISSN :
- 00221198
- Volume :
- 65
- Issue :
- 4
- Database :
- OpenAIRE
- Journal :
- Journal of Forensic Sciences
- Accession number :
- edsair.doi.dedup.....bcb68e79ec3f50349e39b5c5b0f241ef