• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권3호
• /
• pp.881-895
• /
• 2023

Modern image inpainting techniques based on deep learning have achieved remarkable performance, and more and more people are working on repairing more complex and larger missing areas, although this is still challenging, especially for facial image inpainting. For a face image with a huge missing area, there are very few valid pixels available; however, people have an ability to imagine the complete picture in their mind according to their subjective will. It is important to simulate this capability while maintaining the identity features of the face as much as possible. To achieve this goal, we propose a three-stage network model, which we refer to as the identity and structure feature refinement network (ISFRNet). ISFRNet is based on 1) a pre-trained pSp-styleGAN model that generates an extremely realistic face image with rich structural features; 2) a shallow structured network with a small receptive field; and 3) a modified U-net with two encoders and a decoder, which has a large receptive field. We choose structural similarity index (SSIM), peak signal-to-noise ratio (PSNR), L1 Loss and learned perceptual image patch similarity (LPIPS) to evaluate our model. When the missing region is 20%-40%, the above four metric scores of our model are 28.12, 0.942, 0.015 and 0.090, respectively. When the lost area is between 40% and 60%, the metric scores are 23.31, 0.840, 0.053 and 0.177, respectively. Our inpainting network not only guarantees excellent face identity feature recovery but also exhibits state-of-the-art performance compared to other multi-stage refinement models.

• 정보과학회 논문지
• /
• 제41권9호
• /
• pp.666-673
• /
• 2014

본 논문에서는 비제약적 얼굴 데이터 베이스를 위한 확장성 있는 얼굴 인식 방법을 연구하고, 간단한 실험 결과를 소개한다. 기존의 얼굴 인식 연구들은 주로 조명, 얼굴 각도, 표정, 배경 등 제약이 있는 환경에서의 정확도 향상에 초점을 맞추고 있어서 비제약적 얼굴 데이터 베이스에 사용하기에 적합하지 않다. 제안하는 얼굴인식 방법은 비제약적 얼굴 인식을 위한 특징 추출 알고리즘으로, 먼저 지역적 특징이 존재하는 눈, 코, 입과 같이 얼굴의 중요한 특징을 나타내는 영역을 분리한다. 각 얼굴 주요 위치는 고차원의 다중 스케일 국부 이진패턴 히스토그램(Multi-scale LBP histogram) 특징 벡터로 기술된다. 단일 얼굴 주요 위치에 해당하는 다중 스케일 국부 이진패턴 히스토그램 특징 벡터는 주성분 분석법(PCA: Principal Component Analysis)과 선형 판별 분석법(LDA: Linear Discriminant Analysis)의 차원 축소 과정을 통해 저차원 얼굴 특징 벡터를 생성한다. 저차원 얼굴 특징 벡터는 랭크 획득과 Precision at k(p@k) 성능 평가 방법을 이용하여 제안한 알고리즘의 얼굴 인식 성능을 검증한다. 본 연구는 FERET, LFW 및 PubFig83 데이터 베이스를 이용하여 얼굴 인식 실험을 수행하였으며, 제안한 알고리즘을 이용한 얼굴 인식 방법이 기존의 방법보다 향상된 인식성능을 보였다.

• 전자공학회논문지
• /
• 제52권7호
• /
• pp.86-92
• /
• 2015

본 논문에서는 지역적인 에지의 방향 정보와 반응 크기, 주변 화소와의 밝기값 차이를 LDP 코드에 포함함으로써 얼굴 표정 인식률을 향상시킨다. 기존 LDP 코드를 사용하면 LBP에 비해서 영상의 밝기 변화에 덜 민감하고 잡음에 강한 장점을 가진다. 하지만, 밝기 변화가 없는 매끄러운 영역의 정보를 표현하기 어렵고, 배경에 얼굴과 유사한 에지 패턴이 존재하는 경우에는 인식률이 저하되는 문제점이 있다. 따라서 에지 방향 정보를 기반으로 에지 강도 및 밝기값을 추가할 수 있도록 LDP 코드를 개선하고, 인식률을 측정한다.

• 디지털산업정보학회논문지
• /
• 제5권2호
• /
• pp.167-176
• /
• 2009

In this paper, a system that detects driver's drowsiness has been implemented based on the automatic extraction and the tracking of pupils. The research also focuses on the compensation of illumination and reduction of background noises that naturally exist in the driving condition. The system, that is based on the principle of Haar-like feature, automatically collects data from areas of driver's face and eyes among the complex background. Then, it makes decision of driver's drowsiness by using recognition of characteristics of pupils area, detection of pupils, and their movements. The implemented system has been evaluated and verified the practical uses for the prevention of driver's drowsiness.

• 제어로봇시스템학회논문지
• /
• 제16권12호
• /
• pp.1150-1158
• /
• 2010

In this study, the Polynomial-based Radial Basis Function Neural Networks is proposed as one of the recognition part of overall face recognition system that consists of two parts such as the preprocessing part and recognition part. The design methodology and procedure of the proposed pRBFNNs are presented to obtain the solution to high-dimensional pattern recognition problem. First, in preprocessing part, we use a CCD camera to obtain a picture frame in real-time. By using histogram equalization method, we can partially enhance the distorted image influenced by natural as well as artificial illumination. We use an AdaBoost algorithm proposed by Viola and Jones, which is exploited for the detection of facial image area between face and non-facial image area. As the feature extraction algorithm, PCA method is used. In this study, the PCA method, which is a feature extraction algorithm, is used to carry out the dimension reduction of facial image area formed by high-dimensional information. Secondly, we use pRBFNNs to identify the ID by recognizing unique pattern of each person. The proposed pRBFNNs architecture consists of three functional modules such as the condition part, the conclusion part, and the inference part as fuzzy rules formed in 'If-then' format. In the condition part of fuzzy rules, input space is partitioned with Fuzzy C-Means clustering. In the conclusion part of rules, the connection weight of pRBFNNs is represented as three kinds of polynomials such as constant, linear, and quadratic. Coefficients of connection weight identified with back-propagation using gradient descent method. The output of pRBFNNs model is obtained by fuzzy inference method in the inference part of fuzzy rules. The essential design parameters (including learning rate, momentum coefficient and fuzzification coefficient) of the networks are optimized by means of the Particle Swarm Optimization. The proposed pRBFNNs are applied to real-time face recognition system and then demonstrated from the viewpoint of output performance and recognition rate.

• 대한의용생체공학회:의공학회지
• /
• 제44권5호
• /
• pp.329-335
• /
• 2023

Perceived age is defined as age estimated based on physical appearance. Perceived age is an important indicator of the overall health status of the elderly. This is because people who appear older tend to have higher rates of morbidity and mortality than people of the same chronological age. Although perceived age is an important indicator, there is a lack of objective methods to quantify perceived age. In this paper, we construct a quantified perceived age model from face images using a convolutional neural network. The face images are enlarged to super-resolution and the skin, an important feature in perceived age, is made clear. Moreover, through Tanh-polar transformation, the central area of the face occupies a relatively larger area than the boundary area, helping the neural network better recognize facial skin features. The experimental results show mean absolute error (MAE) of 6.59, showing that the proposed model is superior to existing method.

• Journal of Multimedia Information System
• /
• 제5권1호
• /
• pp.15-20
• /
• 2018

There are a lot of problems in the face detection area. One of them is detecting faces by facial features and reducing number of the false negatives and positions. This paper is directed to solve this problem by the proposed triangle method. Also, this paper explans cascades, Haar-like features, AdaBoost, HOG. We propose a scheme using 12-net, 24-net, 48-net to scan images and improve efficiency. Using triangle method for frontal pose, B and B1 methods for other poses in neural networks are proposed.

• International journal of advanced smart convergence
• /
• 제8권4호
• /
• pp.104-112
• /
• 2019

This study proposed a robust detection algorithm. It detects face more stably with respect to changes in light and rotation forthe identification of a face shape. The proposed algorithm uses face shape asinput information in a single camera environment and divides only face area through preprocessing process. However, it is not easy to accurately recognize the face area that is sensitive to lighting changes and has a large degree of freedom, and the error range is large. In this paper, we separated the background and face area using the brightness difference of the two images to increase the recognition rate. The brightness difference between the two images means the difference between the images taken under the bright light and the images taken under the dark light. After separating only the face region, the face shape is recognized by using the self-organization feature map (SOFM) algorithm. SOFM first selects the first top neuron through the learning process. Second, the highest neuron is renewed by competing again between the highest neuron and neighboring neurons through the competition process. Third, the final top neuron is selected by repeating the learning process and the competition process. In addition, the competition will go through a three-step learning process to ensure that the top neurons are updated well among neurons. By using these SOFM neural network algorithms, we intend to implement a stable and robust real-time face shape recognition system in face shape recognition.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2003년도 학술회의 논문집 정보 및 제어부문 B
• /
• pp.1002-1005
• /
• 2003

This paper presents a new template matching method for finding facial feature points. Instead of matching each template to its corresponding feature point separately the present method matches a set of element templates simultaneously. The set of templates can be placed on the space predefined by the geometrical characteristics of human faces. As a result, the search area for template matching is very small compared with a conventional method. This makes the presented method very robust and accurate. Experiment results show that the presented method results in good performance In various illuminance environments and poses.

• 사상체질의학회지
• /
• 제22권2호
• /
• pp.37-47
• /
• 2010

1. Objectives: We aimed to derive constitutional facial characteristics with front and side picture of the people in this study 2. Methods: Through November 2007 to July 2009, we obtained front and side face photograph data of 715 male/female constitution confirmed subjects within the age range of 10-80 from 19 oriental medical facilities in the country. According to sex, we divided the subjects into two groups, real constitution group and non-constitution group. We analyzed significant variables of distance, angle and ratio of facial point through unpaired T-test. 3. Results: There are different significant variables according to Sasang Constitution and even though they are much different between sex, the interpretation of the meaning of those variables are similar with the written characteristics of ancient Writing. The face size of the Taeeumin is bigger than that of the non-Taeeumin, and especially they have long length from mouth to mandible angle. Soeumin have small jaw area, long philtrum and narrow facial feature like egg. Soyangin has developed forehead and glabella, short jaw, long eye side. Taeyangin has developed forehead, weak nose especially side end of nose, long eye side. 4. Conclusions: Sasang Constitutional facial characteristics are different between man and woman but there are much similarity with the facts as known on the interpretation of significant variables characteristics.