• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.3
• /
• pp.1100-1118
• /
• 2021

In a wide range of ML applications, the training data contains privacy-sensitive information that should be kept secure. Training the ML systems by privacy-sensitive data makes the ML model inherent to the data. As the structure of the model has been fine-tuned by training data, the model can be abused for accessing the data by the estimation in a reverse process called model inversion attack (MIA). Although, MIA has been applied to shallow neural network models of recognizers in literature and its threat in privacy violation has been approved, in the case of a deep learning (DL) model, its efficiency was under question. It was due to the complexity of a DL model structure, big number of DL model parameters, the huge size of training data, big number of registered users to a DL model and thereof big number of class labels. This research work first analyses the possibility of MIA on a deep learning model of a recognition system, namely a face recognizer. Second, despite the conventional MIA under the white box scenario of having partial access to the users' non-sensitive information in addition to the model structure, the MIA is implemented on a deep face recognition system by just having the model structure and parameters but not any user information. In this aspect, it is under a semi-white box scenario or in other words a gray-box scenario. The experimental results in targeting five registered users of a CNN-based face recognition system approve the possibility of regeneration of users' face images even for a deep model by MIA under a gray box scenario. Although, for some images the evaluation recognition score is low and the generated images are not easily recognizable, but for some other images the score is high and facial features of the targeted identities are observable. The objective and subjective evaluations demonstrate that privacy cyber-attack by MIA on a deep recognition system not only is feasible but also is a serious threat with increasing alert state in the future as there is considerable potential for integration more advanced ML techniques to MIA.

• Therapeutic Science for Rehabilitation
• /
• v.11 no.4
• /
• pp.69-83
• /
• 2022

Objective : This study aimed to confirm the possibility of the clinical application of oral motor facilitation technique (OMFT) protocol and neuromuscular electrical stimulation (NMES) in patients with Wallenberg syndrome. Methods : One patient with Wallenberg syndrome was treated with OMFT and NMES applied 40 times each, 5 days a week, twice a day for 4 weeks. The Comprehensive Oral-Facial Function Scale (COFFS), Korean-Mann Swallowing Ability Assessment (K-MASA), and Penetration-Aspiration Scale (PAS) were used to compare the changes before and after the intervention. Data analysis was used to compare the score changes before and after the intervention. Results : Orofacial function and swallowing ability improved after the intervention in the individual who participated in this study. Among oral motor functions, relatively greater functional improvement was observed in tongue movement compared to other functions, which was evaluated to the extent that pharyngeal swallowing was possible. Conclusions : Early swallowing rehabilitation using systematic OMFT and NMES of exercise intensity confirmed the possibility of improving oral motor function and dysphagia. In the future, complementary studies on the effects of interventions applying the OMFT and NMES will be needed.

• Archives of Craniofacial Surgery
• /
• v.23 no.4
• /
• pp.171-177
• /
• 2022

Background: The treatment of zygoma complex fractures is of crucial importance in the field of plastic surgery. However, surgical methods to correct zygoma complex fractures, including the number of fixation sites, differ among operators. Although several studies have compared two-point and three-point fixation, no comparative research has yet been conducted on one-point versus two-point fixation using computed tomography scans of surgical results. Therefore, the present study aimed to address this gap in the literature by comparing surgical results between one-point and two-point fixation procedures. Methods: In this study, we randomly selected patients to undergo surgery using one of two surgical methods. We analyzed patients with unilateral zygoma complex fractures unaccompanied by other fractures according to whether they underwent one-point fixation of the zygomaticomaxillary buttress or two-point fixation of the zygomaticomaxillary buttress and the zygomaticofrontal suture. We then made measurements at three points-the zygomaticofrontal suture, inferior orbital wall, and malar height-using 3-month postoperative computed tomography images and performed statistical analyses to compare the results of the two methods. Results: All three measurements (zygomaticofrontal suture, inferior orbital wall, and malar height) showed significant differences (p< 0.05) between one-point and two-point fixation. Highly significant differences were found for the zygomaticofrontal suture and malar height parameters. The difference in the inferior wall measurements was less meaningful, even though it also reached statistical significance. Conclusion: Using three parameters in a statistical analysis of imaging findings, this study demonstrated significant differences in treatment outcomes according to the number of fixations. The results indicate that bone alignment and continuity can be achieved to a greater extent by two-point fixation instead of one-point fixation.

• The Journal of Korean Academy of Prosthodontics
• /
• v.61 no.2
• /
• pp.125-134
• /
• 2023

When the patient with class III malocclusion needs extensive oral rehabilitation due to multiple missing teeth, accurate diagnosis, and careful analysis, such as the patient's occlusal relationship, facial changes, and evaluation of the temporomandibular joint are essential. Orthognathic surgery is often performed for aesthetic improvement, depending on the patient's chief complaint. If it is not possible due to certain circumstances, partial aesthetic improvement can be achieved through minimal elevation of the vertical dimension. As this patient may have unexpected issues, such as temporomandibular joint disorder, oral habits like bruxism, and masticatory muscle tension, it was determined whether the patient could adjust to a reversible temporary removable partial denture. After this, the maxillary implant-supported fixed prostheses and the mandibular fixed prostheses were used to achieve stable posterior support and to partially improve the maxillary anterior esthetics. The patient was satisfied with the results both aesthetically and functionally. The prognosis is expected to be good if regular check-ups are conducted.

• Journal of the Korean Society of Floral Art and Design
• /
• no.41
• /
• pp.45-59
• /
• 2019

The characters that could be connected to diverse cultural contents have formed diverse platforms with the development of digital technology, and the size of the relevant industry and market is rapidly growing. Recently, the utilization of character emoticons for smartphone messenger has been rapidly increased, so that the characters are settled down as a tool for non-verbal communication, on top of drawing attention as an independent area. With the expansion of character market, the importance of design that could give interest and familiarity to consumers is more emphasized. The body proportion of characters includes the implicative and symbolic meanings that could express diverse personalities. Thus, this study examined the body proportion of the characters with the high consumers' preference, and then analyzed the characteristics of formative elements of character design in accordance with the body proportion. In the results of the analysis, the exaggerated form of SD characters in two or three-head figure, and the realistic Real characters in seven or eight-head figure were preferred. For the SD characters, the colors with a high chroma showing the cute and cheerful image were used. For the Real characters, the cubic effect was expressed through the colors with active images and the light and shade of color. Even though the SD characters have limited motions due to the omitted body parts, the facial movements of animation characters are exaggerated while the Real characters describe the realistic and dynamic motions.

• Restorative Dentistry and Endodontics
• /
• v.47 no.1
• /
• pp.3.1-3.9
• /
• 2022

Objectives: This study aimed to assess the impact of a glide-path on the shaping ability of 2 single-file instruments and to compare the centering ability, maintenance of original canal curvatures and area of instrumentation in simulated S-shaped root canals. Materials and Methods: Forty simulated S-shaped root canals were used and were prepared with One Curve (group OC), One G and OC (group GOC), Reciproc Blue (group RB) and R-Pilot and RB (group PRB) and scanned before and after instrumentation. The images were analyzed using AutoCAD. After superimposing the samples, 4 levels (D1, D2, D3, and D4) and 2 angles (Δ1 and Δ2) were established to evaluate the centering ability and modification of the canal curvatures. Then, the area of instrumentation (ΔA) was measured. The data were analyzed using 2-way analysis of variance and Tukey's test for multiple comparisons (p < 0.05). Results: Regarding the centering ability in the apical part (D3, D4), the use of the glide-path yielded better results than the single-file groups. Among the groups at D4, OC showed the worst results (p < 0.05). The OC system removed less material (ΔA) than the RB system, and for Δ1, OC yielded a worse result than RB (p < 0.05). Conclusions: The glide-path improved the centering ability in the apical part of the simulated S-shaped canals. The RB system showed a better centering ability in the apical part and major respect of the canal curvatures compared with OC system.

• Industry Promotion Research
• /
• v.8 no.1
• /
• pp.1-10
• /
• 2023

In this study, a formula (EJ-F101) was prepared to develop a raw material for acne-prone skin improvement using wheat germ extract, and a clinical trial cream was prepared and clinical trials were conducted. As a result of the analysis, when comparing before and after using the product, both the test group and the control group showed significant improvement effects in terms of open comedones, occluded comedones, papules, sebum and oil content in the facial region at 4 weeks after product use, compared to the control group in the test group which showed a more significant improvement effect. As a result of the survey on the efficacy of the product, most items showed higher positive answers in the test product compared to the control product four weeks after the use of the product, and about 43-81% of the study subjects answered positively in the test product, except for the "open surface" item. In addition, for all items related to the usability of the product, about 14-86% of the test group and 38-90% of the control group answered positively at the time point 4 weeks after using the product. As a result of skin safety evaluation, no adverse skin reactions were observed in all subjects of this study. Based on the above results, it is considered that the cream using wheat germ extract is suitable for use on acne-prone skin(non comedogenicity).

• Journal of Korean Medicine Rehabilitation
• /
• v.33 no.4
• /
• pp.125-144
• /
• 2023

Objectives This study aimed to investigate domestic clinical research trends in Shinbaro pharmacopuncture, identify diseases managed using Shinbaro pharmacopuncture, and suggest the direction for future studies to increase its clinical utilization. Methods This study used five steps proposed by Arksey and O'Malley and the PRISMA-extension for scoping reviews checklist. We examined published literature on Shinbaro pharmacopuncture studies reported until June 5, 2023, in the following eight databases (Research Information Sharing Service, Science ON, Oriental Medicine Advanced Searching Integrated System, KMBASE, The Society of Internal Korean Medicine, PubMed, EMBASE, and the Cochrane Library). The search terms used were 'Shinbaro' or 'Sinbaro'. Results A total of 47 studies were included in our analysis. Of these, 37 (78.7%) were interference time series studies. Shinbaro pharmacopuncture was the most frequently used treatment for lumbosacral disease (n=15). In the facial area, ST4 and ST6 were used in five out of six studies, and in the shoulder area, TE14 and LI15 were used in all studies. Nine of the 15 studies in the lumbosacral area used the EX-B2. The other parts mostly used the pressure points. Compared to other pharmacopuncture methods, the treatment effect was similar to that of bee venom, and faster than that of jungsongouhyul. Conclusions This is the first scoping review of Shinbaro pharmacopuncture therapy in South Korea. Studies with a high level of evidence based on sole treatment, large capacities, and standardization of Shinbaro pharmacopuncture need to be conducted to increase its clinical utilization.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2024.01a
• /
• pp.55-58
• /
• 2024

최근 각 기업의 AI 면접시스템 도입이 증가하고 있으며, AI 면접에 대한 실효성 논란 또한 많은 상황이다. 본 논문에서는 AI 면접 과정에서 지원자를 평가하는 방식을 시각, 음성, 자연어처리 3영역에서 구현함으로써, 면접 지원자를 다방면으로 분석 방법론의 적절성에 대해 평가하고자 한다. 첫째, 시각적 측면에서, 면접 지원자의 감정을 인식하기 위해, 합성곱 신경망(CNN) 기법을 활용해, 지원자 얼굴에서 6가지 감정을 인식했으며, 지원자가 카메라를 응시하고 있는지를 시계열로 도출하였다. 이를 통해 지원자가 면접에 임하는 태도와 특히 얼굴에서 드러나는 감정을 분석하는 데 주력했다. 둘째, 시각적 효과만으로 면접자의 태도를 파악하는 데 한계가 있기 때문에, 지원자 음성을 주파수로 환산해 특성을 추출하고, Bidirectional LSTM을 활용해 훈련해 지원자 음성에 따른 6가지 감정을 추출했다. 셋째, 지원자의 발언 내용과 관련해 맥락적 의미를 파악해 지원자의 상태를 파악하기 위해, 음성을 STT(Speech-to-Text) 기법을 이용하여 텍스트로 변환하고, 사용 단어의 빈도를 분석하여 지원자의 언어 습관을 파악했다. 이와 함께, 지원자의 발언 내용에 대한 감정 분석을 위해 KoBERT 모델을 적용했으며, 지원자의 성격, 태도, 직무에 대한 이해도를 파악하기 위해 객관적인 평가지표를 제작하여 적용했다. 논문의 분석 결과 AI 면접의 다면적 평가시스템의 적절성과 관련해, 시각화 부분에서는 상당 부분 정확도가 객관적으로 입증되었다고 판단된다. 음성에서 감정분석 분야는 면접자가 제한된 시간에 모든 유형의 감정을 드러내지 않고, 또 유사한 톤의 말이 진행되다 보니 특정 감정을 나타내는 주파수가 다소 집중되는 현상이 나타났다. 마지막으로 자연어처리 영역은 면접자의 발언에서 나오는 말투, 특정 단어의 빈도수를 넘어, 전체적인 맥락과 느낌을 이해할 수 있는 자연어처리 분석모델의 필요성이 더욱 커졌음을 판단했다.

• Tuberculosis and Respiratory Diseases
• /
• v.59 no.3
• /
• pp.298-305
• /
• 2005

Background : Obstructive sleep apnea syndrome (OSAS) is believed to have multifactorial causes. The major risk factors for OSAS are obesity, narrowed upper airways, and abnormal cranial-facial structures. A genetic basis for OSAS has been also suggested by reports of families with many members affected. This study analyzed the HLA typing in patients with OSAS to determine the possible role of genetics in OSAS. Methods : Twenty-five Korean patients with OSAS (1 woman and 24 men; age range 30-66 years) were enrolled in this study. A diagnosis of OSAS was made using full-night polysomnography. The control group consisted of 200 healthy Korean people. Serologic typing of the HLA-A and B alleles was performed in all patients using a standard lymphocyte microcytotoxicity test. Analysis of the polymorphic second exons of the HLA-DRB1 gene was performed using a polymerase chain reaction-sequence specific oligonucleotide probe. Results : The allele frequency of HLA-A11 was significantly lower in patients with OSAS compared with the controls (p<0.05). The HLA-B allele frequencies in the patients and controls had a similar distribution. Analysis of the HLADRB1 gene polymorphisms showed an increased frequency of DRB1*09 in the OSA patients compared with the controls (p<0.05). When the analysis was performed after dividing the OSAS patients according to the severity of apnea, the allele frequency of HLA-DRB1*08 was significantly higher in the severe OSA patients (apnea index >45) than in the controls (p<0.05). Conclusion : This study revealed an association between OSAS and the HLA-A11 and DRB1*09 alleles as well as association between the disease severity and the HLA-DRB1*08 allele in Korean patients. These results suggest that genetics plays an important role in both the development and the disease severity of OSAS.