• Journal of Digital Convergence
• /
• v.12 no.12
• /
• pp.513-520
• /
• 2014

Since the beginning of history, men have done mimesis and produced illusion and succeeded art and culture instinctually. The subject which mention above included the object which can order and space around that. Perspective which began the Renaissance age was dominant way about understanding space in western history and it made modern visual system. Direction way of space which based perspective is changed as horizontal data included multi-layer structure in digital media age. This character make us possible to represent the space more efficiently. So we must have pay attention the direction way of space based on digital media, because it has meaning to show human value beyond a methodology of visual art culture.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.41 no.1
• /
• pp.38-46
• /
• 2004

Image segmentation is the task which partitions the image into meaningful regions and considered to be one of the most important steps in computer vision and image processing. Image segmentation is also widely used in object-based video compression such as MPEG-4 to extract out the object regions from the given frame. Watershed algorithm is frequently used to obtain the more accurate region boundaries. But, it is well known that the watershed algorithm is extremely sensitive to gradient noise and usually results in oversegmentation. To solve such a problem, we propose an image segmentation method which is robust to noise by using stabilized inverse diffusion equation (SIDE) and is more efficient in segmentation by employing multiresolution approach. In this paper, we apply both the region projection method using labels of adjacent regions and the region merging method based on region adjacency graph (RAG). Experimental results on noisy image show that the oversegmenation is reduced and segmentation efficiency is increased.

• 한국IT서비스학회:학술대회논문집
• /
• 2008.11a
• /
• pp.187-190
• /
• 2008

데이터 스트림이란 제한 없이 끊임없이 흘러 들어오는 일련의 많은 양의 데이터 객체들을 의미하며, 센서 데이터 처리, 인터넷 트래픽 분석, 웹 서버 로그와 같은 다양한 트랜잭션 로그 분석등과 관련된 수많은 응용 분야에 적용 가능하기 때문에 이들을 처리 하기 위해 많은 연구가 진행되었다. 데이트 스트림을 처리하기 위해서는 미리 등록된 질의들(연속 질의)을 새롭게 들어오는 스트림 데이터들로 계산하여 그 결과를 계속적으로 생성하여야 하므로 연속 질의들은 스트림 데이터가 들어올 때마다 반복적으로 수행되며, 데이터 스트림은 매우 빠르게 입력되는 특성을 가지고 있기 때문에 보다 빠르게 질의를 처리하여야만 한다. 본 논문에서는 다수의 조인 연속 질의들이 시스템에 등록되어 있을 때, 이들을 보다 빠르게 처리할 수 있도록 여러 개의 질의에 반복적으로 적용되는 조인 연산들을 공유함으로써 최적의 질의 계획을 생성하는 기법을 제안한다.

• 한국HCI학회:학술대회논문집
• /
• 2007.02a
• /
• pp.744-749
• /
• 2007

물체 추적시스템은 비디오 감시 시스템, 화상회의 시스템과 같은 다양한 비전 응용 분야에서 점점 비중이 높아지고 있다. 이 시스템에서 가장 널리 사용되고 있는 방법 중 하나로 Particle-Filter를 들 수 있다. 하지만, 이 Particle-Filter의 단점은 유사한 여러 물체를 추적할 때에 그 물체들이 겹치거나 사라질 경우 정확한 추적을 하기 어렵다는 것이다. 이 단점을 극복하기 위해 많은 연구가 진행되고 있으며, 본 논문에서는 이 문제를 극복하기 위한 새로운 방법을 제안하고자 한다. 다중 물체 추적에서 빈번히 일어나는 문제는 두 가지로 요약할 수 있는데, 동일한 다중 물체가 부분적으로 엇갈리거나 다른 객체에 완전히 겹친 후 떨어질 때 한 물체를 중복하여 추적하는 문제(merge and split problem)와 이 때 분리되어 추적은 됐지만, 물체를 혼동하여 추적하는 문제(Labeling problem)이다. 본 논문에서는 이 러한 문제들을 풀기 위해 이미지 필드에서 보다 정확한 확률분포를 만들고, 이 확률분포의 신뢰성을 높이기 위해서 물체의 특징정보를 표현하는 몇 가지 방법을 제안한다. 전자의 문제는 두 가지 문제로 나누어 생각해 보았다. 첫째, 복잡환 환경에서의 분포를 찾아내는 것과 둘째, 추적 중인 물체를 잃어버릴 경우 새로운 샘플을 생성함으로써 나누어 보았다. 이 문제 중 첫번째는 K-means 클러스터링을 이용하여 유사한 물체가 주변에 퍼져 있을 때, 하나의 후보 위치가 아닌, K개의 후보 위치들을 만들어 내어 보다 정확한 추적이 가능하게 하였으며, 두 번째 문제는 추적 중인 물체가 다른 커다란 물체에 가려질 경우이다. 이 상황에서 샘플을 생성하는 방법은 지금까지 해왔던 간단한 환경에서의 생성 범위와는 다르게 넓게 해야 생성시켜야 한다. 이 때 샘플링의 수를 늘리지 않으면서, 최대한 정확하게 추적하기 위해서 동영상에서 물체의 모션을 이용한 모션 히스토그램을 얻어내고, 그 정보를 이용하여 샘플을 생성하는 위치를 조절함으로써 이 문제를 풀어 보았다. 그리고, 후자의 문제인 이미지 필드상에서 확률분포의 신뢰성을 높이기 위한 특징 정보는 기존에 많이 사용하던 칼라 히스토그램에 공간정보의 의미를 부여하는 칼라 히스토그램을 분할하는 방법과 SIFT에서 사용하는 방향정보와 크기정보를 사용했다. 이것들을 사용하여 보다 정확한 물체추적시스템을 다음과 같이 제안한다.

• Korean Journal of Remote Sensing
• /
• v.39 no.5_3
• /
• pp.967-978
• /
• 2023

This study proposed and validated a combined segmentation network (CSN) designed to effectively train on multiple drone image datasets and enhance the accuracy of semantic segmentation. CSN shares the entire encoding domain to accommodate the diversity of three drone datasets, while the decoding domains are trained independently. During training, the segmentation accuracy of CSN was lower compared to U-Net and the pyramid scene parsing network (PSPNet) on single datasets because it considers loss values for all dataset simultaneously. However, when applied to domestic autonomous drone images, CSN demonstrated the ability to classify pixels into appropriate classes without requiring additional training, outperforming PSPNet. This research suggests that CSN can serve as a valuable tool for effectively training on diverse drone image datasets and improving object recognition accuracy in new regions.

• The Transactions of the Korea Information Processing Society
• /
• v.5 no.11
• /
• pp.2797-2808
• /
• 1998

Multimedia scenario database system is a read-only multimedia-on-demand system which transfers scenarios representing the display ordering of multimedia objects. A scenario is a graph of multimedia objects and it contains spatial, temporal and contextual information of multimedia data. By structuring multimedia objects as a scenario, it is possible to enforce their display order based on their context. Furthermore, it can provide multiple display paths as well as the sharing of objects between different scenarios. As a result, the multimedia scenario database system can perform the pre-scheduling of multimedia objects, which makes it possible to reorder the transmission order of objects in a scenario. Consequently, the overall system resource such as data buffer and network bandwidth can be highly utilized. In this paper, we discuss the requirements of structuring a scenario to design a scenario database that stores and manages multimedia scenario. Furthermore, we devise and analyze several scheduling policies based on the reordering mechanism for the objects in a scenario.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.27 no.5
• /
• pp.585-597
• /
• 2009

This research aims to develop a registration method to remove the geometric inconsistency between aerial images and LIDAR data acquired from an airborne multi-sensor system. The proposed method mainly includes registration primitives extraction, correspondence establishment, and EOP(Exterior Orientation Parameters) adjustment. As the registration primitives, we extracts planar patches and intersection edges from the LIDAR data and object points and linking edges from the aerial images. The extracted primitives are then categorized into horizontal and vertical ones; and their correspondences are established. These correspondent pairs are incorporated as stochastic constraints into the bundle block adjustment, which finally precisely adjusts the exterior orientation parameters of the images. According to the experimental results from the application of the proposed method to real data, we found that the attitude parameters of EOPs were meaningfully adjusted and the geometric inconsistency of the primitives used for the adjustment is reduced from 2 m to 2 cm before and after the registration. Hence, the results of this research can contribute to data fusion for the high quality 3D spatial information.

• Asia-pacific Journal of Multimedia Services Convergent with Art, Humanities, and Sociology
• /
• v.9 no.11
• /
• pp.1133-1144
• /
• 2019

Recently, the amount of image on the Internet has rapidly increased, due to the advancement of smart devices and various approaches to effective image retrieval have been researched under these situation. Existing image retrieval methods simply detect the objects in a image and carry out image retrieval based on the label of each object. Therefore, the semantic gap occurs between the image desired by a user and the image obtained from the retrieval result. To reduce the semantic gap in image retrievals, we connect the module for multiple objects classification based on deep learning with the module for human behavior classification. And we combine the connected modules with a behavior ontology. That is to say, we propose an image retrieval system considering the relationship between objects by using the combination of deep learning and behavior ontology. We analyzed the experiment results using walking and running data to take into account dynamic behaviors in images. The proposed method can be extended to the study of automatic annotation generation of images that can improve the accuracy of image retrieval results.

• Journal of Internet Computing and Services
• /
• v.9 no.3
• /
• pp.43-50
• /
• 2008

Authenticated Encryption scheme in RFID system is the important issue for ID security. But, implementing authenticated Encryption scheme in RFID systems is not an easy proposition and systems are often delivered for reasons of complexity, limited resources, or implementation, fail to deliver required levels of security. RFID system is so frequently limited by memory, performance (or required number of gates) and by power drain, that lower levels of security are installed than required to protect the information. In this paper, we design a new authenticated encryption scheme based on the EC(Elliptic Curve)'s x-coordinates and scalar operation. Our scheme will be offers enhanced security feature in RFID system with respect to user privacy against illegal attack allowing a ECC point addition and doubling operation.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.23 no.1
• /
• pp.65-70
• /
• 2023

In this paper, we used a deep learning model to detect and segment flame and smoke in real time from fires. To this end, well known U-NET was used to separate and divide the flame and smoke of the fire using multi-class. As a result of learning using the proposed technique, the values of loss error and accuracy are very good at 0.0486 and 0.97996, respectively. The IOU value used in object detection is also very good at 0.849. As a result of predicting fire images that were not used for learning using the learned model, the flame and smoke of fire are well detected and segmented, and smoke color were well distinguished. Proposed method can be used to build fire prediction and detection system.