개인용 얼굴 식별 시스템을 위한 새로운 심화 신경망 구조 = A New Deep Neural Network Architecture For Personal Face Identification Systems

다국어 초록 (Multilingual Abstract)

We investigated personal face identification systems using deep learning networks and we proposed a deep neural network architecture for improving false positive rate. Most of conventional face identification systems have the same number of output nodes as the number of faces that can be identified, and each node is trained to identify one registered face. In this paper, we added an extra node to identify unregistered faces for improving false positive rate(FPR) and accuracy by reducing the propagation of matching probabilities toward other output nodes identifying the registered faces, when unregistered faces are input. The proposed model has been trained with the VGGFace2 dataset. The performance of the proposed model was analyzed in terms of accuracy, precision, FPR, and false negative rate (FNR), and was compared to that of the existing model. According to the performance analysis results, when FNR is 5%, the FPR of the proposed model is improved by about 83% compared to the existing model.

국문 초록 (Abstract)

딥러닝 네트워크를 활용한 개인용 얼굴 식별 시스템을 연구하였으며, 오탐율을 개선하기 위한 새로운 심화 신경망 구조를 제안하였다. 기존 대부분의 딥러닝을 활용한 얼굴 식별 시스템들에서는 식별 가능한 얼굴의 개수와 동일한 출력 노드들이 존재하고 각 출력 노드들은 하나의 등록 얼굴을 식별하도록 학습된다. 본 논문에서는 미등록 얼굴을 식별하기 위한 출력 노드를 추가하였으며 이를 통해서 미등록 얼굴 입력 시에 매칭 확률이 다른 등록 얼굴 식별을 담당하는 출력 노드들로 전파되는 현상을 완화시킴으로써 전체적인 정확도와 오탐율을 개선하였다. 제안된 딥러닝 모델을 VGGFace2 데이터 셋을 활용하여 학습시키고 accuracy, precision, false positive rate (FPR), 그리고 false negative rate (FNR)측면에서 성능을 분석하고 기존 방식과 비교하였다. 성능 분석 결과에 따르면, FNR이 5%일 경우에 제안 방식의 FPR은 기존 방식대비 약 83% 개선됨을 확인할 수 있었다.

참고문헌 (Reference)

• 1 "https://www.robots.ox.ac.uk/~vgg/data/vgg_face2/"
• 2 "https://support.apple.com/en-us/HT213062"
• 3 "https://pytorch.org/vision/main/models/generated/torchvision.models.resnet152.html"
• 4 "https://github.com/timesler/facenet-pytorch"
• 5 Q. Cao, "VGGFace2: A Dataset for Recognising Faces across Pose and Age" 67-74, 2018
• 6 G. B. Huang, "Labeled faces in the wild: A database for studying face recognition in unconstrained environments"
• 7 Janet Hui-wen Hsiao, "Impact of mask use on face recognition: an eye-tracking study" Springer Science and Business Media LLC 7 (7): 1-15, 2022
• 8 KH Teoh, "Face Recognition and Identification using Deep Learning Approach" IOP Publishing 1755 (1755): 1-10, 2020
• 9 Miracle Udurume ; Angela Caliwag ; 임완수 ; 김귀곤, "Emotion Recognition Implementation with Multimodalities of Face, Voice and EEG" 한국정보통신학회 20 (20): 174-180, 2022
• 10 I. Masi, "Deep Face Recognition: A Survey" 471-478, 2018
• 11 J. S. Lim, "An Efficient Recognition Scheme for Obscured Face based on GAN Algorithm" 374-376, 2022
• 12 R. Raj, "A survey on machine learning-based facial recognition algorithm" 1-7, 2022
• 13 Lixiang Li, "A Review of Face Recognition Technology" Institute of Electrical and Electronics Engineers (IEEE) 8 : 139110-139120, 2020
• 14 R. Sharma, "A Review Paper on Facial Recognition Techniques" 617-621, 2021