• Cartoon and Animation Studies
• /
• s.41
• /
• pp.337-356
• /
• 2015

Recently augmented reality contents are variously used in public such as advertisements or exhibits as well as children's books. Therefore, it is certain that the market, development of augmented reality contents, is gradually growing. Those who are the producer of augmented reality may be familiar with the skill where those images are used as a marker which is created by image recognition technique. In case of using image recognition technique, they usually use the augmented reality marker platform from Qualcomm since it is able to recognize self-produced images and 3-dimensional figures at no cost. This study was started when undergraduate students began to use those general techniques in their contents producing process. AR majoring students in Namseoul University applied image recognition technique to 3 AR contents exhibited in Sejong Center. Creating 3 different images, they have registered images at Image Target Manager provided by Vuforia to use as a marker. Moreover, they have modified the image producing method to raise the recognition rate by research. The higher recognition rate brings the more stable use of augmented reality contents. To achieve the satisfied rate, they have compared the elements of color contrast, pattern and etc. in the use of platform. Thus, the effective image creation method has been drawn. This study is aiming to suggest the production of stable contents by recognizing smart devices' limitation and producing educational contents. The purpose of this study is to help practically augmented reality contents developers by illustrating the application of augmented reality contents which are based on image recognition technique and also its effectiveness at the same time.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.05a
• /
• pp.446-450
• /
• 2020

딥러닝 기반의 영상 분석 방법들은 많은 양의 학습 데이터가 필요하며, 학습 데이터 구축에는 많은 시간과 노력이 소요된다. 특히 객체 검출 분야의 경우 영상 내 객체의 위치, 크기, 범주 등의 정보가 모두 필요하여 학습 데이터 구축에 더 많은 어려움이 있으며, 이를 해결하기 위해 최근 이미지 합성기반 데이터 증강에 관한 연구가 활발히 진행되고 있다. 이미지 합성기반 데이터 증강 방법은 배경 영상에 객체를 합성할 때 객체와 배경 영상이 접한 영역에서 아티팩트(Artifact)가 발생하며, 이는 객체 검출 모델이 아티팩트를 객체의 특징으로 모델링하여 검출 성능이 저하되는 원인이 된다. 이러한 문제를 해결하기 위하여 본 논문에서는 양방향 필터 기반의 이미지 합성 방법을 제안하고, 단일 단계 검출의 대표적인 방법인 RetinaNet을 이용하여 이미지 합성기반 데이터 증강 방법의 성능을 분석하였다. 공개 데이터셋에 대한 실험 결과 본 논문에서 사용한 단일 검출 방법 및 데이터 증강 기법을 사용하면 더 적은 양의 증강 데이터로 기존 방법과 동일한 성능을 보여주는 것을 확인하였다.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.12
• /
• pp.167-174
• /
• 2023

This paper attempted to improve the accuracy of the colorectal cancer diagnosis model using image data augmentation in convolutional neural network. Image data augmentation was performed by flipping, rotation, translation, shearing and zooming with basic image manipulation method. This study split 4000 training data and 1000 test data for 5000 image data held, the model is learned by adding 4000 and 8000 images by image data augmentation technique to 4000 training data. The evaluation results showed that the clasification accuracy for 4000, 8000, and 12,000 training data were 85.1%, 87.0%, and 90.2%, respectively, and the improvement effect depending on the increase of image data was confirmed.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2022.07a
• /
• pp.539-540
• /
• 2022

본 논문에서는 Bounding Box가 포함된 증강 이미지 데이터셋을 손쉽게 생성할 수 있는 독립형 GUI 프로그램을 제안한다. 본 논문의 연구를 통하여 직관적인 마우스 클릭 동작만으로 적은 수의 이미지 파일과 annotation 파일로부터 필요한 만큼의 증강 이미지 데이터셋을 짧은 시간 내에 생성하고, 다양한 아키텍처의 학습용 이미지 데이터셋 증강에 적용할 수 있다.

• Cartoon and Animation Studies
• /
• s.33
• /
• pp.503-529
• /
• 2013

The commemorative AR postage stamps which are the world first presented at The YEOSU EXPO 2012 has had meaning of communicating with future in this present from a convergence that the most analog medium is using now and that the AR is cutting edge of digital technology. The AR stamps printed 10 kind out of 33 commemorative stamps. These have great significance that is artistic value than that is world first. The applied AR images are not only expressed 3D real images but also artic represented and signifying each stamp images from visualized creativity process, and build 'new art space' that is new concept between on real(analog) and virtual(digital). This study analyzes meaning of images and then makes concept of AR contents design. The processing is designed and considered the meaning of architectures and environments, and the regional specific feature of the Yeosu with surrealistic graphic concept. The 10 of deducted images were expressed after AR coding such as visual arts. This study realized markerless 3D image tracking AR stamps and deducted research result are; the first, it was able to figure out how to realize AR in the process of registering the reference images, coordinating transformation, and hybriding AR on the stamps for the mobile devices. The second, it was able to be seeked a possibility of new virtual exhibition space. The third, it was able to know possibility of satisfaction of immersing with visual formativeness and usability with informativity.

• Journal of the Korea Computer Graphics Society
• /
• v.29 no.3
• /
• pp.79-84
• /
• 2023

Image generation models have been applied in various fields to overcome data sparsity, time and cost issues. However, it has limitations in generating images from regular pattern images and detecting defects in such data. In this paper, we verified the feasibility of the image generation model to generate pattern images and applied it to data augmentation for defect detection of OLED panels. The data required to train an OLED defect detection model is difficult to obtain due to the high cost of OLED panels. Therefore, even if the data set is obtained, it is necessary to define and classify various defect types. This paper introduces an OLED panel defect data acquisition system that acquires a hypothetical data set and augments the data with an image generation model. In addition, the difficulty of generating pattern images in the diffusion model is identified and a possibility is proposed, and the limitations of data augmentation and defect detection data augmentation using the image generation model are improved.

• Journal of Digital Convergence
• /
• v.12 no.5
• /
• pp.221-229
• /
• 2014

The study considers the meaning of the composite images created when users capture present images over past images tagged location data in using the mobile augmented reality technology. The composite image through the location-based augmented reality technology is the result of matching the same location data between present images users are capturing and past images captured already. It is the new composite images that contain two different narratives-current and past in the same space and in real-time. We developed the mobile application implemented augmented reality technology and analysed the process that users create multi-layered narrative in the middle of capturing present image through augmented reality module. In addition, through the comparison with similar studies and applications of the augmented reality, we found that the key to give the multi-narrative in the composite images is the user's participation to put its personal intentions in real-time capturing process. In further development, we'll be able to utilize the application in order that users easily create multi-layered narrative composite image using cultural and personal records.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.11a
• /
• pp.43-45
• /
• 2022

본 논문은 황반변성 진단 모델 개발을 위해 안저 사진을 이용한 MobileNet2 전이학습 모델 개발과 안정적인 모델 성능을 위한 이미지 증강 방법 및 모델 성능 향상을 위한 파라미터 조정 방법을 제안한다. 보유하고 있는 이미지의 수가 매우 적다고 하더라도 적절한 전이학습 모델을 사용하고 이미지 증강 시 증강 방법과 증강한 이미지와 정상 이미지와의 비율을 적절히 고려할 경우 충분히 안정적인 결과를 얻어낼 수 있다. 또한 파라미터 조정을 통해서 성능 향상을 도모할 수 있다

• Journal of IKEEE
• /
• v.26 no.2
• /
• pp.176-185
• /
• 2022

Data augmentation is general approach to solve overfitting of machine learning models by applying various data transformations and distortions to dataset. However, when data augmentation is applied in GAN-based model, which is deep learning image generation model, data transformation and distortion are reflected in the generated image, then the generated image quality decrease. To prevent this problem called augmentation leak, we propose a scheme that can prevent augmentation leak regardless of the type and number of augmentations. Specifically, we analyze the conditions of augmentation leak occurrence by type and implement auxiliary augmentation task classifier that can prevent augmentation leak. Through experiments, we show that the proposed technique prevents augmentation leak in the GAN model, and as a result improves the quality of the generated image. We also demonstrate the superiority of the proposed scheme through ablation study and comparison with other representative augmentation leak prevention technique.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• fall
• /
• pp.283-286
• /
• 2021

4차 산업혁명을 맞이해 최근 산업 및 기술 영역에서는 인공지능을 이용한 생산력 향상, 자동화 등 딥러닝의 보편화가 빠르게 진행되고 있다. 또한, 딥러닝의 성능을 도출하기 위해서는 수많은 양의 학습용 데이터가 필요하며 그 데이터의 양은 딥러닝 모델의 성능과 정비례한다. 이에 본 작품은 최신형 영상처리 Library인 Albumentations를 이용하여 영상처리 알고리즘을 이용하여 이미지를 증강하고, 이미지 데이터 크롤링 기능을 통해 Web에서 영상 데이터를 수집을 자동화하며, Label Pix를 연동하여 수집한 데이터를 라벨링 한다. 더 나아가 라벨링 된 데이터의 증강까지 포함하여 다양한 증강 자동화를 한 인터페이스에 집적시켜 딥러닝 모델을 생성할 때 데이터 수집과 전처리를 수월하게 한다. 또한, Neural Net 기반의 AdaIN Transfer를 이용하여 이미지를 개별적으로 학습하지 않고 Real time으로 이미지의 스타일을 옮겨올 수 있도록 하여 그림 데이터의 부족 현상을 해결한다.