• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2007.10a
• /
• pp.127-130
• /
• 2007

Skin detection research is important role in the 3G of mobile phone for video telephony and security system by using face recognition. We propose skin detection algorithm as preprocessing to the face recognition, and use YCbCr color space. In existing skin detection algorithm using CbCr, skin colors that is brightened by camera flash or sunlight at outdoor in images doesn't acknowledged the skin region. In order to detect skin region accuracy into any circumstance, this paper proposes 3-channel skin detection algorithm.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2013.10a
• /
• pp.83-86
• /
• 2013

In this paper, by using the Morphing, if the size of the eyes of both sides are not the same, we propose a method to correct the size of eyes area. First, by using the Haar-like feature from a input image that is input, to detect the shape of the eyes and face. After inverting the left and right eye region of one of the shape of the eyes detected sets the correspondence between the second with a line to control the shape of the eyes detected using eyes that is detected with canny edge, in the previous step. To the Warping to match the correspondence was then set in the previous step, an area of each eye. Then, I merge the image which merged in the eye area is detected from the original image. As a result, a system result of the experiment in the test image and face image seen from the front, the proposed, prove to be more efficient than a method of keying the size of the eye only.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• v.9 no.1
• /
• pp.1149-1152
• /
• 2005

Detecting faces in images with complex backgrounds is a difficult task. Our approach, which obtains state of the art results, is based on a generative neural network model: the Constrained Generative Model (CGM). To detect side view faces and to decrease the number of false alarms, a conditional mixture of networks is used. To decrease the computational time cost, a fast search algorithm is proposed. The level of performance reached, in terms of detection accuracy and processing time, allows to apply this detector to a real word application: the indexation of face images on the Web.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.20 no.11
• /
• pp.2054-2060
• /
• 2016

As the aging society is accelerated, the population with dementia grows significantly faster. Because there is no complete cure of dementia, it is of great importance to detect the disease at a very early stage and prevent its fast aggravation through evaluation of MCI(Mild Cognitive Impairment) which happens before dementia. The current assessment of MCI conducts only in the form of hand-written data and by 1:1 face to face communication between the subjects and the examiners. Due to this, both examiners' fatigue rate and evaluation error rate increase. Further, there are some limitations in utilizing and analyzing data collected. Against the problems, therefore, there is a need to develop a computerized system by which MCI is assessed, and the results are saved and analyzed. This paper presents the development of such a system enabling MCI evaluation, its related data collection and analysis.

• Journal of IKEEE
• /
• v.26 no.3
• /
• pp.430-436
• /
• 2022

In this paper, we propose an intelligent camera for multiple body temperature detection. The proposed camera is composed of optical(4056*3040) and thermal(640*480), which detects abnormal symptoms by analyzing a person's facial expression and body temperature from the acquired image. The optical and thermal imaging cameras are operated simultaneously and detect an object in the optical image, in which the facial region and expression analysis are calculated from the object. Additionally, the calculated coordinate values from the optical image facial region are applied to the thermal image, also the maximum temperature is measured from the region and displayed on the screen. Abnormal symptom detection is determined by using the analyzed three facial expressions(neutral, happy, sadness) and body temperature values. In order to evaluate the performance of the proposed camera, the optical image processing part is tested on Caltech, WIDER FACE, and CK+ datasets for three algorithms(object detection, facial region detection, and expression analysis). Experimental results have shown 91%, 91%, and 84% accuracy scores each.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.10
• /
• pp.437-446
• /
• 2022

Text classification task from Natural Language Processing (NLP) combined with state-of-the-art (SOTA) Machine Learning (ML) and Deep Learning (DL) algorithms as the core engine is widely used to detect and classify voice phishing call transcripts. While numerous studies on the classification of voice phishing call transcripts are being conducted and demonstrated good performances, with the increase of non-face-to-face financial transactions, there is still the need for improvement using the latest NLP technologies. This paper conducts a benchmarking of Korean voice phishing detection performances of the pre-trained Korean language model KoBERT, against multiple other SOTA algorithms based on the classification of related transcripts from the labeled Korean voice phishing dataset called KorCCVi. The results of the experiments reveal that the classification accuracy on a test set of the KoBERT model outperforms the performances of all other models with an accuracy score of 99.60%.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.5
• /
• pp.1-8
• /
• 2021

Coronavirus disease 2019 (COVID-19) has affected the world seriously. Every person is required for wearing a mask properly in a public area to prevent spreading the virus. However, many people are not wearing a mask properly. In this paper, we propose an efficient mask detection system. In our proposed system, we first detect the faces of input images using YOLOv5 and classify them as the one of three scene complexity classes (Simple, Moderate, and Complex) based on the number of detected faces. After that, the image is fed into the Faster-RCNN with the one of three ResNet (ResNet-18, 50, and 101) as backbone network depending on the scene complexity for detecting the face area and identifying whether the person is wearing the mask properly or not. We evaluated our proposed system using public mask detection datasets. The results show that our proposed system outperforms other models.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.6
• /
• pp.43-52
• /
• 2022

We can sense somebody's feeling fatigue, which means that fatigue can be detected through sensing human biometric signals. Numerous researches for assessing fatigue are mostly focused on diagnosing the edge of disease-level fatigue. In this study, we adapt quantitative analysis approaches for estimating qualitative data, and propose video analysis models for measuring fatigue state. Proposed three deep-learning based classification models selectively include stages of video analysis: object detection, feature extraction and time-series frame analysis algorithms to evaluate each stage's effect toward dividing the state of fatigue. Using frontal face videos collected from various fatigue situations, our CNN model shows 0.67 accuracy, which means that we empirically show the video analysis models can meaningfully detect fatigue state. Also we suggest the way of model adaptation when training and validating video data for classifying fatigue.

• Science of Emotion and Sensibility
• /
• v.25 no.4
• /
• pp.107-118
• /
• 2022

Human can rapidly detect and deal with dangerous elements in their environment, and they generally manifest as attentional bias toward threat. Past studies have reported that this attentional bias is affected by anxiety level. Other studies, however, have argued that children and adolescents show attentional bias to threatening stimuli, regardless of their anxiety levels. Few studies directly have compared the two age groups in terms of attentional bias to threat, and furthermore, most previous studies have focused on attentional capture and the early stages of attention, without investigating further attentional holding by the stimuli. In this study, we investigated both attentional bias patterns (attentional capture and holding) with respect to negative emotional stimulus in neurotypical adults and adolescents. The effects of anxiety level on attentional bias were also examined. The results obtained for adult participants showed that abrupt onset of a distractor delayed attentional capture to the target, regardless of distractor type (angry or neutral faces), while it had no effect on attention holding. In adolescents, on the other hand, only the angry face distractor resulted in longer reaction time for detecting a target. Regarding anxiety, state anxiety revealed a significant positive correlation with attentional capture to a face distractor in adult participants but not in adolescents. Overall, this is the first study to investigate developmental tendencies of attentional bias to negative facial emotion in both adults and adolescents, providing novel evidence on attentional bias to threats at different ages. Our results can be applied to understanding the attentional mechanisms in people with emotion-related developmental disorders, as well as typical development.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.43 no.3 s.309
• /
• pp.17-24
• /
• 2006

In this paper, an algorithm for anchor frame detection in news video is proposed, which consists of four steps. In the first step, the cumulative histogram method is used to detect shot boundaries in order to segment a news video into video shots. In the second step, skin color information is used to detect face regions in each shot boundary. In the third step, color information of upper body regions is used to extract anchor object, which produces candidate anchor frames. Then, from the candidate anchor frames, a graph-theoretic cluster analysis algorithm is utilized to classify the news video into anchor-person frames and non-anchor frames. Experiment results have shown the effectiveness of the proposed algorithm.