Ship detection in synthetic aperture radar (SAR) images is an important application in marine monitoring for the military and civilian domains. Over the past decade, object detection has achieved significant progress with the development of convolutional neural networks (CNNs) and lot of labeled databases. However, due to difficulty in collecting and labeling SAR images, it is still a challenging task to solve SAR ship detection CNNs. To overcome the problem, some methods have employed conventional data augmentation techniques such as flipping, cropping, and affine transformation, but it is insufficient to achieve robust performance to handle a wide variety of types of ships. In this paper, we present a novel and effective approach for deep SAR ship detection, that exploits label-rich Electro-Optical (EO) images. The proposed method consists of two components: a data augmentation network and a ship detection network. First, we train the data augmentation network based on conditional generative adversarial network (cGAN), which aims to generate additional SAR images from EO images. Since it is trained using unpaired EO and SAR images, we impose the cycle-consistency loss to preserve the structural information while translating the characteristics of the images. After training the data augmentation network, we leverage the augmented dataset constituted with real and translated SAR images to train the ship detection network. The experimental results include qualitative evaluation of the translated SAR images and the comparison of detection performance of the networks, trained with non-augmented and augmented dataset, which demonstrates the effectiveness of the proposed framework.

In this paper, we present an entry-level COVID-19 stand-alone digitial signage player (CoSiP) which performs not only conventional digital signage functionalities but also fever detection, face mask wearing detection, and KI-pass QR code checking. The overall design of CoSiP is proposed, and implementation of a temperature checking algorithm using a low cost thermal sensor is elaborately presented. Through experiments over datasets and against a developed CoSiP device, it is shown that the fever detection, face mask wearing detection, KI-pass QR code checking as well as signage functionalities of the proposed CoSiP work properly and reliably.

Although image compression is one of the essential technologies to transmit image data on a variety of surveillance and mobile healthcare applications, it causes unnecessary compression artifacts such as blocking and ringing artifacts by the lossy compression in the limited network bandwidth. Recently, image restoration methods using convolutional neural network (CNN) show the significant improvement of image quality from the compressed images. In this paper, we propose Image Denoising Convolutional Neural Networks (IDCNN) to reduce the compression artifacts for the purpose of improving the performance of object classification. In order to evaluate the classification accuracy, we used the ImageNet test dataset consisting of 50,000 natural images and measured the classification performance in terms of Top-1 and Top-5 accuracy. Experimental results show that the proposed IDCNN can improve Top-1 and Top-5 accuracy as high as 2.46% and 2.42%, respectively.

Image-based gender classification and age estimation of human are classic problems in computer vision. Most of researches in this field focus just only one task of either gender classification or age estimation and most of the reported methods for each task focus on accuracy performance and are not computationally light. Thus, running both tasks together simultaneously on low cost mobile or embedded systems with limited cpu processing speed and memory capacity are practically prohibited. In this paper, we propose a novel light-weight gender classification and age estimation method based on ensemble multitasking deep learning with light-weight processing neural network architecture, which processes both gender classification and age estimation simultaneously and in real-time even for embedded systems. Through experiments over various well-known datasets, it is shown that the proposed method performs comparably to the state-of-the-art gender classification and/or age estimation methods with respect to accuracy and runs fast enough (average 14fps) on a Jestson Nano embedded board.

Few-shot learning is a sub-area of machine learning problems, which aims to classify target images that only contain a few labeled samples for training. As a representative few-shot learning method, the Prototypical network has been received much attention due to its simplicity and promising results. However, the Prototypical network uses the sample mean of samples from the same class as the prototypes of that class, which easily results in learning uncharacteristic features in the low-data scenery. In this study, we propose to use local descriptors (i.e., patches along the channel within feature maps) from the same class to explicitly obtain more representative prototypes for Prototypical Network so that significant intra-class feature information can be maintained and thus improving the classification performance on few-shot learning tasks. Experimental results on various benchmark datasets including mini-ImageNet, CUB-200-2011, and tiered-ImageNet show that the proposed method can learn more discriminative intra-class features by the local descriptors and obtain more generic prototype representations under the few-shot setting.

When communication such as light, radio wave, or magnetic field is used underwater, the communication distance is very short, so sound waves are mainly used. However, by combining the strengths of each medium and communicating, stable communication may be possible. Underwater multi-media communication requires a protocol stack that supports it, which is very complex. To this end, this paper proposes a standard protocol stack and modeling technique to enable easy protocol stack modeling for the purpose. In fact, in this paper, a random model was created and analyzed through the proposal of modeling elements and similarity measurement methods, and as a result, it was analyzed that it was very helpful in creating a new model based on a standard model.

Visual Object Tracking is known as the most fundamental problem in the field of computer vision. Object tracking localize the region of target object with bounding box in the video. In this paper, a custom CNN is created to extract object feature that has strong and various information. This network was constructed as a Siamese network for use as a feature extractor. The input images are passed convolution block composed of a bottleneck layers, and features are emphasized. The feature map of the target object and the search area, extracted from the Siamese network, was input as a local proposal network. Estimate the object area using the feature map. The performance of the tracking algorithm was evaluated using the OTB2013 dataset. Success Plot and Precision Plot were used as evaluation matrix. As a result of the experiment, 0.611 in Success Plot and 0.831 in Precision Plot were achieved.

In this paper, time series prediction method of photovoltaic power is introduced using seasonal autoregressive integrated moving average (SARIMA). In order to obtain the best fitting model by a time series method in the absence of an environmental sensor, this research was used data below 50% of cloud cover. Three samples were extracted by time intervals from the raw data. After that, the best fitting models were derived from mean absolute percentage error (MAPE) with the minimum akaike information criterion (AIC) or beysian information criterion (BIC). They are SARIMA (1,0,0)(0,2,2)14, SARIMA (1,0,0)(0,2,2)28, SARIMA (2,0,3)(1,2,2)55. Generally parameter of model derived from BIC was lower than AIC. SARIMA (2,0,3)(1,2,2)55, unlike other models, was drawn by AIC. And the performance of models obtained by SARIMA was compared. MAPE value was affected by the seasonal period of the sample. It is estimated that long seasonal period samples include atmosphere irregularity. Consequently using 1 hour or 30 minutes interval sample is able to be helpful for prediction accuracy improvement.

This study suggests a variety of ways to prevent narcotic crimes, recurrence of such crimes, rehabilitation, and improve the public awareness of using the narcotics by adopting cultural contents, highlighting how dangerous, harmful and they are easily accessible, as well as the reality in Korea nowadays, where it cannot any longer be seen as a drug-free state. Precedent research, the actual case studies, and interviews followed by the "Musical Mission" explain that anyone can easily become a criminal or family to the criminals, and the adolescent is the age group that experience the most rapid increase in initiating such illicit drugs by using cultural content to help prevent and improve awareness among the public to expose the deadly effect of the substances. More efforts are required to remove the frame of criminals off the active drug users, as they are those who only need to be cared and are the same members of our society after all by demonstrating new ways to change such perspectives. I wish that it will become a social mandatory for those who produce the social contents to urge more support for the drug users through TV advertisements, dramas, movies, documentaries, and musicals etc.

Based on the information framework and regulatory focus theory, this paper verifies the impact on the advertising effect when the matching effect forms between the video advertising content type (Information Type, Entertainment Type) and the information framework (Promotional, Prevention) and the regulatory focus (Promotion Orientation, Prevention Orientation) in social media. In order to verify, after selecting of the experimental subjects and finishing the prior investigation of the experimental operation, four video advertisements and formal experiments were performed. The experimental results showed that first, the content type affects the advertising effect of the video and the information framework; when the information video matches the preventive information framework, it will have a more obvious positive effect on the advertising effect; when the entertainment video matches the promotion information framework, it will have a more obvious positive effect on the advertising effect. Second, with the two matches of video content type and information framework, the entertainment-oriented video has a more obvious positive effect for the information-oriented video audience; the above results enrich user created content research and improve the quality and communication efficiency of video in social media.

Korean web drama are undergoing changes before and after the popularization of the OTT platform. The previous web drama were mainly aimed at teenagers, with school content as a genre feature, and the active casting of new idols and the use of B-grade emotions as production characteristics. After the popularization of OTT represented by YouTube and Netflix, there are signs of significant change. This study examined the hyperrealistic content characteristics of the web drama , which can be the center of the change. Through this, it was considered that has a greatly changed aspect from previous web dramas through stereoscopic typification, emphasis on detailed expression, active use of experiential narrative characteristics, and a new media creator-centered production method.

Based on existing research, the psychological state of Chinese students has become a very significant issue that needs to be resolved. In addition to paying attention to the daily life and study of Chinese students, the psychological problems of Chinese students are also worthy of attention. At the same time, if the existing psychological problems are not resolved in time, serious consequences may result. Based on the ART(Attention Restoration Theory) theory, this article uses VR (Virtual Reality) content as a medium, uses 3D modeling software to build a healing scene that helps Chinese students improve their psychological and emotional state, and presents it in a VR device. To achieve the purpose of improving the psychological and emotional state of Chinese students. According to experimental tests, the VR recovery scene constructed by this method can help improve the psychological mood of Chinese international students who already have subliminal depression. The results of independent sample T-tests after data analysis experiments show that after the intervention of the experiment, the depression of the experimental group is significantly improved compared to the control group. It is proved that the method in this study is effective for the mentality and emotion of Chinese international students who have subliminal depression. There is a significant improvement effect.

In order to convert Korean folk paintings called Minhwa, a part of traditional Korean heritage, into DBs, it is necessary to design a classification system suitable for the characteristics of folk paintings. A classification system and the generating of unique codes are required to classify and save them. To realize this, a basic classification system was created by listing objects depicted in folk paintings, and keywords were extracted by reclassifying them for each object. In order to assign a unique code to each piece, we organize the English names of each Minhwa since the English names of the folk painting contain the names of objects. The code name is extracted by applying the order of nouns and consonant priority rules in English names and attaching five Arabic numerals. These codes are later assigned to each image file stored in the database and are input together with the keyword. The Minhwa DB constructed in this way enables storage and search centered on objects and keywords and the intuitive inferring of the type of object from the code name.