• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권12호
• /
• pp.4763-4775
• /
• 2020

This research aims to develop a deep learning-based method that automatically detects and segments the defects on railroad surfaces to reduce the cost of visual inspection of the railroad. We developed our segmentation model by modifying a fully convolutional network model [1], a well-known segmentation model used for machine learning, to detect and segment railroad surface defects. The data used in this research are images of the railroad surface with one or more defect regions. Railroad images were cropped to a suitable size, considering the long height and relatively narrow width of the images. They were also normalized based on the variance and mean of the data images. Using these images, the suggested model was trained to segment the defect regions. The proposed method showed promising results in the segmentation of defects. We consider that the proposed method can facilitate decision-making about railroad maintenance, and potentially be applied for other analyses.

• Journal of the Korean Wood Science and Technology
• /
• 제47권3호
• /
• pp.265-276
• /
• 2019

In our previous study, the LeNet3 model successfully classified images from the transverse surfaces of five Korean softwood species (cedar, cypress, Korean pine, Korean red pine, and larch). However, a practical limitation exists in our system stemming from the nature of the training images obtained from the transverse plane of the wood species. In real-world applications, it is necessary to utilize images from the longitudinal surfaces of lumber. Thus, we improved our model by training it with images from the longitudinal and transverse surfaces of lumber. Because the longitudinal surface has complex but less distinguishable features than the transverse surface, the classification performance of the LeNet3 model decreases when we include images from the longitudinal surfaces of the five Korean softwood species. To remedy this situation, we adopt ensemble methods that can enhance the classification performance. Herein, we investigated the use of ensemble models from the LeNet and MiniVGGNet models to automatically classify the transverse and longitudinal surfaces of the five Korean softwoods. Experimentally, the best classification performance was achieved via an ensemble model comprising the LeNet2, LeNet3, and MiniVGGNet4 models trained using input images of $128{\times}128{\times}3pixels$ via the averaging method. The ensemble model showed an F1 score greater than 0.98. The classification performance for the longitudinal surfaces of Korean pine and Korean red pine was significantly improved by the ensemble model compared to individual convolutional neural network models such as LeNet3.

• 한국정보디스플레이학회:학술대회논문집
• /
• 한국정보디스플레이학회 2006년도 6th International Meeting on Information Display
• /
• pp.1569-1574
• /
• 2006

We have studied elemental image compatibility between integral imaging(II) and parallax generation(PG) system for three-dimensional display. The elemental images of PG can be obtained by recombination of the elemental images picked up in II system. The theoretical verification and the experimental results show that the elemental images of PG are in correspondence with the elemental images of II system with proper transformation conditions.

• 한국의류학회지
• /
• 제41권6호
• /
• pp.1085-1097
• /
• 2017

In the era of new media, images hold an important position as episteme to express and convey ideas. Fashion films provide dynamic and unique fashion images, differentiated from prior fashion media as a representational tool for showing a realistic fashion image only; consequently, their production and spread are increasing rapidly as a new fashion media. This study identifies the meaning and type of fashion images in fashion films based on the concept of Deleuze's image that help discover distinctive characteristics of fashion films as a new fashion media of an expressive tool. Literature research was conducted on new media, concepts and types of images by Deleuze to analyze types of new media images. According to research, fashion image in fashion film is defined as a fashion event; consequently, three types of new media images are derived. As the result of the empirical study, fashion images in fashion films are classified by images of realistic movement, variable time, and virtual experience. The results of the consideration show that fashion films expressed fashion through temporality and narrative, senses, and diegesis. Fashion images of new media in fashion films portray fashion as a process that transcends reality and imagination.

• 복식
• /
• 제66권6호
• /
• pp.158-177
• /
• 2016

In recent years, K-beauty including Korean cosmetics and beauty care tips, is becoming popular in various Asian countries such as China and Vietnam along, with the popularity of the Korean wave. Color images are a highly effective tool in establishing image development strategies in the process of developing cosmetic brands. Surveys on the color images of K-beauty perceived by foreigners need to be preceded for the development of differentiated images and the establishment of management strategies regarding K-beauty. Therefore, the purpose of this study was to examine the color images perceived by Asian consumers about K-beauty, and come up with measures to facilitate K-beauty. To this end, this study selected the two countries, China and Vietnam that show high levels of acceptance of K-beauty, and compared the color images of K-beauty perceived by female consumers in these countries. Then, the study visually suggested these images with color palettes, and compared differences in the perception of color images according to demographic characteristics. Beijing and Shanghai showed similar degrees of perception in most color images of K-beauty whereas Hanoi showed a lower perception level. K-beauty color images were classified into 6 groups: feminine, natural, elegant, modern, sensual, active, and popular, which represent symbolic images of K-beauty.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2003년도 Proceedings of ACRS 2003 ISRS
• /
• pp.525-527
• /
• 2003

Recently, the spatial resolution of earth observation satellites is significantly increased to a few meters. Such high spatial resolution images definitely will provide lots of information for detail-thirsty remote sensing users. However, it is more difficult to develop automated image algorithms for automated image feature extraction and pattern recognition. In this study, we propose a two-stage procedure to extract road information from high resolution satellite images. At first stage, a watershed segmentation technique is developed to classify the image into various regions. Then, a knowledge is built for road and used to extract the road regions. In this study, we use panchromatic and multi-spectral images of the IKONOS satellite as test dataset. The experiment result shows that the proposed technique can generate suitable and meaningful road objects from high spatial resolution satellite images. Apparently, misclassified regions such as parking lots are recognized as road needed further refinement in future research.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권9호
• /
• pp.3298-3321
• /
• 2021

A lot of data hiding schemes have been proposed to embed secret data in the plain cover images or compressed images of various formats, including JPEG, AMBTC, VQ, etc. In this paper, we propose a production process of mosaic images based on three regular images of coffee beans. A primary image is first mimicked by the process to produce a mosaic cover image. A two-layer steganography is applied to hide secret data in the mosaic image. Based on the low visual quality of the mosaic cover image, its PSNR value can be improved about 1.5 dB after embedding 3 bpp. This is achieved by leveraging the newly proposed polarized search mask and the concepts of strong embedding and weak embedding. Applying steganography to the mosaic cover images is a completely new idea and it is promising.

• Journal of information and communication convergence engineering
• /
• 제22권1호
• /
• pp.70-79
• /
• 2024

Bad weather conditions such as haze lead to a significant lack of visibility in images, which can affect the functioning and reliability of image processing systems. Accordingly, various single-image dehazing (SID) methods have recently been proposed. Existing SID methods have introduced effective visibility improvement algorithms, but they do not reflect the image's perspective, and thus have limitations that distort the sky area and nearby objects. This study proposes a new SID method that reflects the sense of space by defining the correlation between image brightness and haze. The proposed method defines the haze intensity by calculating the airlight brightness deviation and sets the weight factor of the depth map by classifying images based on the defined haze intensity into images with a large sense of space, images with high intensity, and general images. Consequently, it emphasizes the contrast of nearby images where haze is present and naturally smooths the sky region to preserve the image's perspective.

• 도시과학
• /
• 제6권2호
• /
• pp.17-27
• /
• 2017

최근, 하천의 수질오염 및 악취발생으로 민원이 증가하여 하천환경개선에 큰 관심이 모아지고 있다. 본 연구의 목적은 하수 유입부에 대해 무인항공기(UAV)를 활용하여 RGB 및 적외 열 영상을 획득하고 하천제방 정비 계획 및 하천 오염 현황의 모니터링을 위한 응용성을 검토하였다. 특히, 하천 인근 공장에서 배출되는 폐수를 적외 열 영상으로 검출하여 폐수의 전파를 모니터링하였다. 또한 하천 제방 정비대상 지역과 인근지역에 대한 RGB영상을 SfM(Structure from Motion)기반 영상 해석을 통해 고정밀 3차원 모형을 제작하고 정확성을 검토하였다. 연구결과, UAV영상을 활용, 폐수유입에 따른 하천의 온도변화를 감지하여 수질오염의 유입부 및 전파 현상을 모니터링 할 수 있었다. 또한 고정밀 3차원 모델(수치지형도, 정사영상)을 제작, 정확성을 검토하고 하천의 제방정비를 위한 정밀 3차원 정보 및 식생 피복정보를 도출할 수 있었다.

• International Journal of Contents
• /
• 제8권3호
• /
• pp.83-93
• /
• 2012

In this paper, we propose a novel method that automatically generates real character images to familiarize existing OCR systems with new fonts. At first, we generate synthetic character images using a simple degradation model. The synthetic data is used to train an OCR engine, and the trained OCR is used to recognize and label real character images that are segmented from ideal document images. Since the OCR engine is unable to recognize accurately all real character images, a substring matching method is employed to fix wrongly labeled characters by comparing two strings; one is the string grouped by recognized characters in an ideal document image, and the other is the ordered string of characters which we are considering to train and recognize. Based on our method, we build a system that automatically generates 2350 most common Korean and 117 alphanumeric characters from new fonts. The ideal document images used in the system are postal envelope images with characters printed in ascending order of their codes. The proposed system achieved a labeling accuracy of 99%. Therefore, we believe that our system is effective in facilitating the generation of numerous character samples to enhance the recognition rate of existing OCR systems for fonts that have never been trained.