• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.39 no.3
• /
• pp.141-148
• /
• 2021

This paper presents a method to detect zebra-crossing using deep learning which combines SegNet and ResNet. For the blind, a safe crossing system is important to know exactly where the zebra-crossings are. Zebra-crossing detection by deep learning can be a good solution to this problem and robotic vision-based assistive technologies sprung up over the past few years, which focused on specific scene objects using monocular detectors. These traditional methods have achieved significant results with relatively long processing times, and enhanced the zebra-crossing perception to a large extent. However, running all detectors jointly incurs a long latency and becomes computationally prohibitive on wearable embedded systems. In this paper, we propose a model for fast and stable segmentation of zebra-crossing from captured images. The model is improved based on a combination of SegNet and ResNet and consists of three steps. First, the input image is subsampled to extract image features and the convolutional neural network of ResNet is modified to make it the new encoder. Second, through the SegNet original up-sampling network, the abstract features are restored to the original image size. Finally, the method classifies all pixels and calculates the accuracy of each pixel. The experimental results prove the efficiency of the modified semantic segmentation algorithm with a relatively high computing speed.

• KIPS Transactions on Computer and Communication Systems
• /
• v.11 no.10
• /
• pp.323-332
• /
• 2022

As various services using AI technology are being developed, much attention is being paid to AI service production. Recently, AI technology is acknowledged as one of ICT services, a lot of research is being conducted for general-purpose AI service production. In this paper, I describe the research results in terms of systems for AI service production, focusing on the distribution and production of machine learning models, which are the final steps of general machine learning development procedures. Three different Ubuntu systems were built, and experiments were conducted on the system, using data from 2017 validation COCO dataset in combination of different AI models (RFCN, SSD-Mobilenet) and different communication methods (gRPC, REST) to request and perform AI services through Tensorflow serving. Through various experiments, it was found that the type of AI model has a greater influence on AI service inference time than AI machine communication method, and in the case of object detection AI service, the number and complexity of objects in the image are more affected than the file size of the image to be detected. In addition, it was confirmed that if the AI service is performed remotely rather than locally, even if it is a machine with good performance, it takes more time to infer the AI service than if it is performed locally. Through the results of this study, it is expected that system design suitable for service goals, AI model development, and efficient AI service production will be possible.

• Journal of the Korean Regional Science Association
• /
• v.38 no.2
• /
• pp.3-20
• /
• 2022

The purpose of this study is to explore the heterogeneity of compensating wage differentials for outdoor workers, under the threat of climate change and heatwave, by region and by wage quantile. This study conducted Oaxaca-Blinder decomposition, multiple regression analysis by region, and unconditional quantile regression analysis using the Korean Working Conditions Survey, which provides individual-level information on the working environment and worker's characteristics. The implications derived from the results of the study are as follows: For most variables, the endowment effect and the price effect were greater for indoor workers, while experience and gender played a role in narrowing the wage gap; The compensating wage differentials for outdoor workers were confirmed to be 2.4% nationwide, depending on the region however, the compensating wage differentials varied from 5 times of national average to nothing statistically significant; The higher the wage quantile, the greater the compensating wage differentials for outdoor workers, and statistically significant monetary compensation was not identified for some low-level outdoor workers. This study is meaningful as an early study that revealed the heterogeneity of compensating wage differentials for outdoor workers and suggested further research on the topic.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2020.06a
• /
• pp.177-177
• /
• 2020

새만금 간척지는 1989년 '새만금간척사업'이 발표된 이래 1991년에 방조제 공사를 시작으로 2006년 물막기 공사를 완료하였고, 2009년 이후 방조제를 완공하여 현재는 40,100ha의 간척지가 조성되었다. 하지만 간척지는 장기적으로 환경에 악영향을 미치며 특히 간척으로 인한 해안경관의 가치 상실, 생태계 파괴 및 교란에 따른 변화가 크기 때문에 이에 대한 대책이 필요하다. 적극적인 식생의 도입이 그 대책의 하나가 될 수 있는데 이는 자연적으로 식물천이를 유도하고 동물들을 유인하여 생태적으로 건강한 환경을 조성하는데 기여할 수 있다. 따라서 본 연구에서는 새만금간척지의 대표적인 수목생육 제한 요인으로 판단되는 염분, 배수 및 통기성에 대한 시험연구를 진행하고자 '09~'11까지 초기에 김제광활에서 2ha의 시험포를 운영하였고, '12부터는 군산옥구에서 2ha의 시험포를 운영하고 있다. 수목을 심을 포지를 구획하기 전에 토양의 염분값을 낮추기 위해서 군산옥구의 경우, 2ha의 시험포를 포함한 전체 묘목장 부지(100ha) 중에서 일부구간(38ha)을 '09 부터 자연강우 담수제염을 통해서 제염을 진행하였다. 이를 통해 초기 염분 값이 18ds/m~20ds/m에서 8ds/m~10ds/m로 낮아졌다. 이렇게 수목이 자랄 수 있는 정도의 제염이 이루어진 이후에는 배수 및 통기성 부분에 초점을 맞추어 포지를 구획하였다. 단지1의 경우, 사전제염작업을 하지 않고, 5M간격의 암거를 설치하였고, 단지2의 경우, 사전강우제염을 진행하고, 10M간격의 암거를 설치하였다. 단지3의 경우, 사전강우제염을 진행하고 그 위에 0.4M의 준설토를 성토하여 포지를 조성하고 일부구간에 5M, 10M암거를 설치하였다. 2년 간의 수목 생존률 및 생장 모니터링을 통해서 각 단지에서의 수목생장의 적합성의 정도를 간접적으로 판단할 수 있었다. 단지3이 암거설치 간격 및 설치유무에 관계없이 수목이 자라는데 가장 적합했다. 다음으로 단지2, 단지1 순으로 나타났다. 염분 값을 낮추기 위한 자연강우 담수제염의 경우에는 문제가 되지 않지만 배수 및 통기성 개선을 위한 암거설치의 경우는 새만금 전체 식재구역에 적용하기에는 경제성이 떨어지므로 배수 및 통기성을 위한 별도의 방안이 요구된다. 또한 각 단지에서의 수목 생장모니터링을 지속적으로 시행하여 보다 장기적인 측면에서의 식재기반 조성기법을 고려할 필요가 있다.

• The Korean Journal of Franchise Management
• /
• v.3 no.1
• /
• pp.26-45
• /
• 2012

Market players have realized the importance of brand power and tried to develop effective marketing programs which focus on consumer's brand experience. This study aims to investigate brand experience and brand personality effect on brand attitude which is overall consumer's faith toward brands and repurchase intention in food-franchised by Structural Equation Model. As results, both brand experience and brand personality affect brand attitude although brand experience has more influence than brand personality. As consumers show positive brand experience and attitude, repurchase intention is higher. Brand attitude plays a mediation role in the relation of brand experience and personality, and repurchase intention. Also brand experience shows more influence than others on repurchase intention.

• Economic Analysis
• /
• v.24 no.3
• /
• pp.1-36
• /
• 2018

When it comes to explaining the relationship between inventory investment and business fluctuations, the production smoothing theory and the stock-out avoidance theory take contradictory stances. Decision-making related to inventory investments of corporations is thought to be influenced by both motives, but the relative sizes or directions of their respective influences can differ depending upon the phase of the business cycle. Against this backdrop, this paper differs from existing studies in that it theoretically tests the relative significances of the production smoothing and stock-out avoidance motives in the inventory investment dynamics, while placing its analytical focus on determining the existence and patterns of the asymmetric dynamics of inventory investment over the business cycle phases. To this end this paper sets up a non-linear model that is expanded from the existing linear inventory investment model, and checks whether its predictive power is better than that of the existing model. The results of analysis confirm the nature of the asymmetric dynamics of inventory investment over the business cycle phases. A stock-out avoidance motive appears but there is no significant production smoothing motive in boom times. In downturns, in contrast, the stock-out avoidance motive is insignificant, but a quality of asymmetric dynamics in which changes in inventory cause the deepening of recessions, due to the non-convexity of production costs proposed by Ramey (1991), is detected. This paper confirms that a model considering the asymmetric dynamics of inventory investment can have better predictive power than one that does not consider it, through within-sample and out-of-sample predictions and various predictive power tests. These research results are expected to be useful for economic forecasting, through their enhancement of the understandings of the inventory investment dynamics and of the nature of its business cycle destabilization.

• Phonetics and Speech Sciences
• /
• v.15 no.3
• /
• pp.75-82
• /
• 2023

Automatic speech recognition (ASR) has been revolutionized with deep learning-based approaches, among which self-supervised learning methods have proven to be particularly effective. In this study, we aim to enhance the performance of OpenAI's Whisper model, a multilingual ASR system on the Korean language. Whisper was pretrained on a large corpus (around 680,000 hours) of web speech data and has demonstrated strong recognition performance for major languages. However, it faces challenges in recognizing languages such as Korean, which is not major language while training. We address this issue by fine-tuning the Whisper model with an additional dataset comprising about 1,000 hours of Korean speech. We also compare its performance against a Transformer model that was trained from scratch using the same dataset. Our results indicate that fine-tuning the Whisper model significantly improved its Korean speech recognition capabilities in terms of character error rate (CER). Specifically, the performance improved with increasing model size. However, the Whisper model's performance on English deteriorated post fine-tuning, emphasizing the need for further research to develop robust multilingual models. Our study demonstrates the potential of utilizing a fine-tuned Whisper model for Korean ASR applications. Future work will focus on multilingual recognition and optimization for real-time inference.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.33 no.6
• /
• pp.1099-1110
• /
• 2023

The recognition system of autonomous driving and robot navigation performs vision work such as object recognition, tracking, and lane detection after multi-sensor fusion to improve performance. Currently, research on a deep learning model based on the fusion of a camera and a lidar sensor is being actively conducted. However, deep learning models are vulnerable to adversarial attacks through modulation of input data. Attacks on the existing multi-sensor-based autonomous driving recognition system are focused on inducing obstacle detection by lowering the confidence score of the object recognition model.However, there is a limitation that an attack is possible only in the target model. In the case of attacks on the sensor fusion stage, errors in vision work after fusion can be cascaded, and this risk needs to be considered. In addition, an attack on LIDAR's point cloud data, which is difficult to judge visually, makes it difficult to determine whether it is an attack. In this study, image scaling-based camera-lidar We propose an attack method that reduces the accuracy of LCCNet, a fusion model (camera-LiDAR calibration model). The proposed method is to perform a scaling attack on the point of the input lidar. As a result of conducting an attack performance experiment by size with a scaling algorithm, an average of more than 77% of fusion errors were caused.

• Journal of the Korea Computer Graphics Society
• /
• v.30 no.3
• /
• pp.81-86
• /
• 2024

The Image Signal Processor (ISP) converts RAW images captured by the camera sensor into user-preferred sRGB images. While RAW images contain more meaningful information for image processing than sRGB images, RAW images are rarely shared due to their large sizes. Moreover, the actual ISP process of a camera is not disclosed, making it difficult to model the inverse process. Consequently, research on learning the conversion between sRGB and RAW has been conducted. Recently, the ParamISP[1] model, which directly incorporates camera parameters (exposure time, sensitivity, aperture size, and focal length) to mimic the operations of a real camera ISP, has been proposed by advancing the simple network structures. However, existing studies, including ParamISP[1], have limitations in modeling the camera ISP as they do not consider the degradation caused by lens shading, optical aberration, and lens distortion, which limits the restoration performance. This study introduces Positional Encoding to enable the camera ISP neural network to better handle degradations caused by lens. The proposed positional encoding method is suitable for camera ISP neural networks that learn by dividing the image into patches. By reflecting the spatial context of the image, it allows for more precise image restoration compared to existing models.

• Applied Microscopy
• /
• v.39 no.4
• /
• pp.333-340
• /
• 2009

In the present study, ultrastructural features of the mesophyll tissue have been investigated in Crassulacean acid metabolism (CAM)-performing succulent Orostachys. A large central vacuole and numerous small vacuoles in the peripheral cytoplasm were characterized at the subcellular level in both developing and mature mesophyll cells. The most notable feature was the invagination of vacuolar membranes into the secondary vacuoles or multivesicular bodies. In many cases, tens of single, membrane-bound secondary vacuoles of various sizes were found to be formed within the central vacuole. multivesicular bodies containing numerous small vesicles were also distributed in the cytoplasm but were better developed within the central vacuole. Occasionally, electron-dense prevacuolar compartments, directly attached to structures appearing to be small vacuoles, were also detected in the cytoplasm. One or more huge central vacuoles were frequently observed in cells undergoing differentiation and maturation. Consistent with the known occurrence of morphologically distinct vacuoles within different tissues, two types of vacuoles, one representing lytic vacuoles and the other, most likely protein storage vacuoles, were noted frequently within Orostachys mesophyll. The two types coexisted in mature vegetative cells but did not merge during the study. Nevertheless, the coexistence of two distinct vacuole types in maturing cells implies the presence of more than one mechanism for vacuolar solute sorting in these species. The vacuolar membrane is known to be unique among the intracellular compartments for having different channels and/or pumps to maintain its function. In CAM plants, the vacuole is a very important organelle that regulates malic acid diurnal fluctuation to a large extent. The membrane invagination seen in Orostachys mesophyll likely plays a significant role in survival under the physiological drought conditions in which these Orostachys occur; by increasing to such a large vacuolar volume, the mesophyll cells are able to retain enormous amounts of acid when needed. Furthermore, the mesophyll cells are able to attain their large sizes with less energy expenditure in order to regulate the large degree of diurnal fluctuation of organic acid that occurs within the vacuoles of Orostachys.