Lightweight face recognition models, as one of the most popular and long-standing topics in the field of computer vision, has achieved vigorous development and has been widely used in many real-world applications due to fewer number of parameters, lower floating-point operations, and smaller model size. However, few surveys reviewed lightweight models and reimplemented these lightweight models by using the same calculating resource and training dataset. In this survey article, we present a comprehensive review about the recent research advances on the end-to-end efficient lightweight face recognition models and reimplement several of the most popular models. To start with, we introduce the overview of face recognition with lightweight models. Then, based on the construction of models, we categorize the lightweight models into: (1) artificially designing lightweight FR models, (2) pruned models to face recognition, (3) efficient automatic neural network architecture design based on neural architecture searching, (4) Knowledge distillation and (5) low-rank decomposition. As an example, we also introduce the SqueezeFaceNet and EfficientFaceNet by pruning SqueezeNet and EfficientNet. Additionally, we reimplement and present a detailed performance comparison of different lightweight models on the nine different test benchmarks. At last, the challenges and future works are provided. There are three main contributions in our survey: firstly, the categorized lightweight models can be conveniently identified so that we can explore new lightweight models for face recognition; secondly, the comprehensive performance comparisons are carried out so that ones can choose models when a state-of-the-art end-to-end face recognition system is deployed on mobile devices; thirdly, the challenges and future trends are stated to inspire our future works.

The Republic of Korea has implemented an obligatory vaccination on major livestock breeds to prevent and eradicate nationwide foot and mouth disease (FMD) since the end of 2010. The government has executed massive serological survey to check the immune level of various herds after vaccination, and seropositive rates against FMD virus (FMDV) structural proteins can be measured to assess FMD immunity level. The purpose of this study is to investigate the FMDV serological level of every cattle breeding farm in the country and to determine whether there is a significant difference between groups classified by time, age, and management authority. A total of 5,781 serum samples was collected in 18 cattle breeding farms from 2020 to 2021, and the seropositive rates were measured using PrioCHECK FMDV Type O ELISA kit. Firstly, the cattle breeding farms were classified by which they are managed: the central government, the local government, and the private agency. Every management authority had a seropositive rate of 99.5% or higher. Secondly, the samples were divided into 6 to 12 months old, 12 to 24 months old, and 24 months or more. The 6 to 12 months old group in 2020 showed a significantly low seropositive rate of 98.1%, but it was improved by implementing the enhanced vaccination policy from 2021 to 100%. In summary, there are considerably high seropositive rates including all groups with time, age, and by which they are managed, which means the FMD vaccination in cattle breeding farms is well-managed.

목적 : 본 연구는 한국어판 Eating Assessment Tool-10(K-EAT-10) 연하장애 선별 검사를 활용하여 지역사회 거주 노인의 연하장애 유병률 및 특성을 확인하기 위해 진행하였다. 연구방법 : 경기도 내 일개 보건소에 내소하는 정상 인지 수준의 지역사회 거주 노인 89명을 K-EAT-10과 연하장애 설문지를 통해 연하장애 증상을 평가하였다. K-EAT-10로 유병률을 확인하였으며, K-EAT-10 3점을 기준으로 연하장애 정상군과 위험군을 나누어 두 그룹 간 차이 및 특성을 Mann-Whitney U로 분석하 였다. 결과 : K-EAT-10 총점 3점 이상의 연하장애 유병률은 24.7%(22명)이었다. 연하장애 증상이 있다고 자각하는 대상자 12명 중 7명은 실제로 연하장애 위험군에 속하였다. 정상군과 위험군 사이에는 연하장 애 증상 및 심리적 요소에서 유의미한 차이가 나타났다. K-EAT-10 항목 별로는 먹을 때 사레 걸림, 목에 음식물이 걸린 느낌, 알약 삼킬 때의 어려움이 주요한 증상으로 나타났다. 결론 : 본 연구를 통하여 일반 노화 과정의 연하장애 증상이 없거나 문제를 호소하지 않는 지역사회 거주 노인들도 내재되어 있는 연하장애 증상 확인을 위한 선제적 검사와 관리가 필요함을 알 수 있었다. 이에 지역사회 전문가들은 지역사회 거주 노인의 연하장애에 대한 관심을 갖고 그에 맞는 종합적인 관리 및 중재를 통한 질 향상 노력을 기울여야 할 것이다.

선박의 안전과 해양사고 예방을 위해서는 선박이 안전한 운항을 담보할 수 있어야 한다. 선박은 선박검사를 통하여 안전운항 적부를 논하게 되며, 검사 규 정의 근간은 IMO 국제해사협약과 각 나라의 국내법이 된다. 우리나라 선박안 전 및 검사제도 법령은 일제 강점기를 거친 역사로 인하여 일본의 그것과 매우 흡사하나 잘 사용하지 않는 용어 수정, 제도 보완 등의 과제가 남아 있다. 또한, 우리는 아직도 해외의 선진법제를 도입하고 있는 처지에 있어 해운 선진국의 위치에 있는 미국과 영국의 법제를 연구하는 것은 우리나라 관련 법제를 정비 하는데 큰 도움이 될 것이다. 이러한 이유로 불문법 체계를 가진 미국과 영국 그리고 우리나라와 같이 성문법 체계를 가진 일본의 선박안전 및 검사 법령을 고찰하였다. 미국과 영국은 불문법 체계를 가지나, 성문법 국가인 우리나라에 비하여 뒤 지지 않을 만큼 제정법화 되어 있으며 통합적인 법 체계를 구성하면서도 세분 화된 검사대상을 규정하여 보다 효과적으로 이행하고 있다. 이에 반하여 우리 나라 선박검사 관련 법령은 여러 기술고시에 산재하여 있는 등 상대적으로 복 잡한 구성을 취하고 있어 수범자로 하여금 규범 준수의 어려움이 있다. 따라서, 선박안전법 및 해양환경관리법 하위의 복잡하고 관리가 어려운 기술고시 및 기 준을 간소화하거나 통합하는 것이 필요하다. 일본의 경우, 법령의 명칭과 용어 사용에 있어서 우리와 매우 유사한 체계를 가지고 있어 구조적인 보완점이나 시사점을 도출하기에는 어려운 사실이 있다. 그러나, 선박안전 및 검사제도 법령을 위반할 경우 벌칙이나 과태료 조항을 관 련 법 조문 뒤에 명기하고 있는 형식을 참조하여 선박검사 관계법령에 한해서 라도 법 조문에 대한 시인성과 직관성을 높이는 방안을 고려할 필요가 있다. 선박검사와 관련하여 국제항해에 종사하는 선박은 국제해사협약을 적용하도 록 하고 국내항해에 종사하는 선박은 각 나라의 특징과 실정에 맞도록 국내법 을 제정하여 적용하는 검사규범 이원화를 고려할 필요가 있다. 또한, 획일화된 규정 제정보다는 특별히 안전관리가 필요한 선박별로 규정을 강화하는 검사규 정 집중화도 요구된다.

This study analyzes the importance and implementation of countermeasures to reduce safety accidents that may occur during fishing in coastal gillnet fishing. Safety accidents that may occur in coastal gillnet fishing were classified into 56 measurement items in six areas: slips and trips, contact with machinery, falls, struck by object, musculoskeletal symptoms, and electrical shock. 35 fishermen belonging to coastal gillnet associations in Tongyeong and Namhae were selected as samples, and the importance and implementation of the proposed measurement items were investigated in the field. In addition, the results were evaluated visually by IPA analysis. As a result of the analysis, the importance of struck by object was the highest, and the implementation of slips and trips was the highest. On the other hand, prevention of musculoskeletal damage was found to be the lowest in both importance and implementation. The area with the most significant difference in importance and implementation by the classified area for measurement items was struck by object, and the area with the minor difference was slips and trips. Among the measurement items, the importance is highly recognized, but there are items with low implementation. To improve their level of implementation, countermeasures such as continuous education, economic support, and structural change of fishing boats should be prepared; moreover, policy support activities for implementation should be carried out.

In this study, we aimed to explore whether eating alone is associated with mental health conditions in Korean adolescents. The data of 2,012 Korean adolescents aged 12-18 years were obtained from the Korea National Health and Nutrition Examination Survey 2015–2019. Participants were classified into three groups based on the frequency of eating alone: none (all meals with others); 1 meal/day alone; and 2 meals/day alone. Mental health conditions were assessed based on stress recognition, depressive symptoms, and suicidal ideation. Multivariable logistic regressions were employed to calculate the adjusted odds ratios (AORs) and 95% confidence intervals (CIs) of poor mental health conditions according to the frequency of eating alone. Adolescents who ate 2 meals/day alone had higher odds of stress recognition (AOR: 2.65, 95% CI: 1.94- 3.63), depressive symptoms (AOR: 2.55, 95% CI: 1.47-4.42), and suicidal ideation (AOR: 2.53, 95% CI: 1.05-6.08) than those who ate all their meals with others. In addition, having breakfast or dinner alone increased the odds of stress recognition. Considering the continuous increase in the social phenomenon of eating alone, nutritional educations are needed to develop adolescents' ability to choose more nutritionally balanced and healthy meals when eating alone.

We present the deep homogeneous 𝑈𝐵𝑉𝑅𝐼 photometric data of 135,071 stars down to 𝑉 ∼ 23 mag and 𝐼 ∼ 22 mag toward the Carina Nebula. These stars are cross-matched with those from the previous surveys in the X-ray, near-infrared, and mid-infrared wavelengths as well as the Gaia Early Data Release 3 (EDR3). This master catalog allows us to select reliable members and determine the fundamental parameters distance, size, stellar density of stellar clusters in this star-forming region. We revisit the reddening toward the nebula using the optical and the near-infrared colors of early-type stars. The foreground reddening [𝐸(𝐵 −𝑉) 𝑓 𝑔] is determined to be 0.35 ± 0.02, and it seems to follow the standard reddening law. On the other hand, the total-toselective extinction ratio of the intracluster medium (𝑅𝑉,𝑐𝑙) decreases from the central region (Trumpler 14 and 16, 𝑅𝑉,𝑐𝑙 ∼ 4.5) to the northern region (Trumpler 15, 𝑅𝑉,𝑐𝑙 ∼ 3.4). It implies that the central region is more dusty than the northern region. We find that the distance modulus of the Carina Nebula to be 11.9 ± 0.3 mag (𝑑 = 2.4 ± 0.35 kpc) using a zero-age main-sequence fitting method, which is in good agreement with that derived from the Gaia EDR3 parallaxes. We also present the catalog of 3,331 pre-main-sequence (PMS) members and 14,974 PMS candidates down to 𝑉 ∼ 22 mag based on spectrophotometric properties of young stars at infrared, optical, and X-ray wavelengths. From the spatial distribution of PMS members and PMS candidates, we confirm that the member selection is very reliable down to faint stars. Our data will have a legacy value for follow-up studies with different scientific purposes.