• The Transactions of the Korea Information Processing Society
• /
• v.6 no.7
• /
• pp.1789-1804
• /
• 1999

This paper presents a simulation-based system for verification and validation(V&V) of design implication of the Visual Real-time Object Model which is produced by existing object's behavior design method. This system can simulate the dynamic interactions using the executable Ada simulation machine, and can detect various logical and temporal problems in the visual real-time object model prior to the real implementation of the application systems. Also, the system can generate the Ada prototype code from the validated specification. This system is implemented by Visual C++ version 4.2. For simulation, this system is using the Ada language because Ada's real-time expression capabilities such as concurrent processes, rendezvous, temporal behavior expression, and etc, are competent compared to other languages. This work contributes to a tightly coupling of methodology-based visual models and formal-based simulation systems, and also contributes to a realization of automated specification V&V.

• Journal of Broadcast Engineering
• /
• v.25 no.6
• /
• pp.1004-1012
• /
• 2020

We present a novel content-preserving seam estimation algorithm for real-time high-resolution video stitching. Seam estimation is one of the fundamental steps in image/video stitching. It is to minimize visual artifacts in the transition areas between images. Typical seam estimation algorithms are based on optimization methods that demand intensive computations and large memory. The algorithms, however, often fail to avoid objects and results in cropped or duplicated objects. They also lack temporal consistency and induce flickering between frames. Hence, we propose an efficient and temporarily-consistent seam estimation algorithm that utilizes a straight line. The proposed method also uses convolutional neural network-based instance segmentation to locate seam at out-of-objects. Experimental results demonstrate that the proposed method produces visually plausible stitched videos with minimal visual artifacts in real-time.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.47 no.3
• /
• pp.51-58
• /
• 2010

Since Augmented Reality (AR) enables users to experience computer-generated content embedded in real environments, AR agents can be visualized among physical objects in the environments where the users exist, and directly interact with them in real-time. We model perceptual attention for autonomous agents in AR environments where virtual and physical objects coexist. Since such AR agents must adaptively perceive and attend to surrounded objects relevant to their goals, our model allows the agents to determine currently visible objects from the description of what virtual and physical objects are configured in the camera's viewing area. A degree of attention is assigned to each perceived object based on its relevance to achieve agents' goals. The agents can focus on a reduced set of perceived objects with respect to the estimated degree of attention. To demonstrate the effectiveness of our approach, we implemented an animated character that was overlaid over a miniature version of campus and that attended to buildings relevant to their goals. Experiments showed that our model could reduce the character's perceptual loads even when surroundings change.

• Journal of Internet Computing and Services
• /
• v.25 no.1
• /
• pp.9-15
• /
• 2024

As virtual reality technology became popular, a high demand emerged for natural and efficient interaction with the virtual environment. Mid-air manipulation is one of the solutions to such needs, letting a user manipulate a virtual object in a 3D virtual space. In this paper, we focus on manipulating a remote virtual object while visually displaying the object and providing tactile information on the object's weight. We developed two types of wearable interfaces that can provide cutaneous or vibrotactile feedback on the virtual object weight to the user's fingertips. Human perception of the remote virtual object weight during manipulation was evaluated by conducting a psychophysics experiment. The results indicate a significant effect of haptic feedback on the perceived weight of the virtual object during manipulation.

• Journal of KIISE
• /
• v.42 no.1
• /
• pp.138-143
• /
• 2015

In this paper we develop a novel human action recognition system based on communication between an ear-mounted Bluetooth camera and an action recognition server to aid scene recognition for the blind. First, if the blind capture an image of a specific location using the ear-mounted camera, the captured image is transmitted to the recognition server using a smartphone that is synchronized with the camera. The recognition server sequentially performs human detection, object detection and action recognition by analyzing human poses. The recognized action information is retransmitted to the smartphone and the user can hear the action information through the text-to-speech (TTS). Experimental results using the proposed system showed a 60.7% action recognition performance on the test data captured in indoor and outdoor environments.

• Journal of KIISE
• /
• v.43 no.7
• /
• pp.795-800
• /
• 2016

In image classification, recent CNNs compete with human performance. However, there are limitations in more general recognition. Herein we deal with indoor images that contain too much information to be directly processed and require information reduction before recognition. To reduce the amount of data processing, typically variational inference or variational Bayesian methods are suggested for object detection. However, these methods suffer from the difficulty of marginalizing over the given space. In this study, we propose an image-text integrated recognition system using active vision based on Spatial Transformer Networks. The system attempts to efficiently sample a partial region of a given image for a given language information. Our experimental results demonstrate a significant improvement over traditional approaches. We also discuss the results of qualitative analysis of sampled images, model characteristics, and its limitations.

• Proceedings of the Korean Information Science Society Conference
• /
• 2007.06c
• /
• pp.460-465
• /
• 2007

영상 내 사물들의 카테고리를 인식하는 연구는 시각적 영상처리와 연관된 다양한 분야에서 활발히 진행되고 있다. 객체 범주화(Object Categorization)는 가정과 같은 실내에서 책상, 의자, 컵, 주전자 등의 다양한 사물들을 구분하여 인식하는데 중요한 역할을 할 수 있다. 본 논문에서는 최근 영상 내 객체들의 카테고리 분석을 위해 연구된 PLSA를 기반으로 특이점에 가중치를 부여하여, 보다 유사한 카테고리 간에 인식 성능을 향상시키는 접근법에 대하여 연구하였다. PLSA는 문서기반의 정보검색 분야로부터 소개된 기법으로, 약한 수준의 비감독 방법임에도 불구하고 인상적인 인식성능을 보여준다. 그러나 비슷한 특징점 분포를 보이는 유사한 카테고리 간의 객체 카테고리 인식에 대해서는 비교적 낮은 성능을 보인다. 본 연구에서는 카테고리간의 비교실험을 통해 각 특징점에 대하여 가중치를 부여한 PLSA를 적용하여 유사한 객체 간의 카테고리 인식 가능성을 살펴보았다. 실험에서는 기존의 PLSA 기법과 제안한 가중치를 부여 PLSA 기법을 각각 적용하여 그 성능을 비교하였다. 본 연구에서는 기존 PLSA 기법에서는 비교적 낮은 인식률을 보인 유사한 카테고리 인식에 대하여 실험 결과를 통해 가중치를 부여한 PLSA 기법이 보다 향상된 성능을 보임을 확인하였다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2006.11a
• /
• pp.59-62
• /
• 2006

본 논문에서는 OpenGL Rendering을 이용한 모델기반 3D 다시점 영상의 객체 구현을 위한 구성과 각 모듈에 적용되는 알고리즘에 대해 중점적으로 연구하였다. 한 장의 텍스쳐 이미지와 깊이 맵(Depth Map)을 가지고 다시점 객체를 생성하기 위해, 먼저 깊이 정보의 전처리 과정을 거친다. 전처리 된 깊이 정보는 OpenGL상에서의 일정 간격의 꼭지점(Vertex) 정보로 샘플링 된다. 샘플링 된 꼭지점 정보는 깊이 정보를 z값으로 가지는 3차원 공간 좌표상의 점이다. 이 꼭지점 정보를 기반으로 텍스쳐 맵핑 (texture mapping)을 위한 폴리곤(polygon)을 구성하기 위해 딜루이니 삼각화(Delaunay Triangulations) 알고리즘이 적용되었다. 이렇게 구성된 폴리곤 위에 텍스쳐 이미지를 맵핑하여 OpenGL의 좌표 연산을 통해 시점을 자유롭게 조정할 수 있는 객체를 만들었다. 제한된 하나의 이미지와 깊이 정보만을 가지고 좀 더 넓은 범위의 시점을 가지는 다시점 객체를 생성하기 위해, 새로운 꼭지점을 생성하여 폴리곤을 확장시켜 기존보다 더 넓은 시점을 확보할 수 있었다. 또한 렌더링된 모델의 경계 영역 부분의 깊이정보 평활화를 통해 시각적인 개선을 이룰 수 있었다.

• Journal of the Korea Society of Computer and Information
• /
• v.13 no.3
• /
• pp.1-13
• /
• 2008

Gestwicki et. al. proposed the technique that detects and predicts the structure of object diagrams from class diagrams to get the improved drawing of object graphs. Their approach finds two predictable structures - recursive clusters and leaf clusters from class diagrams. Their approach is not practical to be applied to real programs, because the starting point is class diagram. In this work, we improve and implement their technique to apply to Java classes. We obtain the structure and relationships of classes and clusters extracted from class files. This information can be provided and used as a template of object graphs for Java execution visualization tools like JIVE.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.47 no.6
• /
• pp.124-130
• /
• 2010

Current traffic guidance system offers the information of traffic light using a speaker to everybody including the blinds without object certification. Therefore, the current system gives stress by noisy sound to normal person. It is difficult for the blind to hear informations such as location, current traffic sign light status, and weather information, because of the vehicle noise and surroundings noise. If the blind or sight disabled person who has the RFID tag or has a white stick with RFID tag approaches to street crossing, RFID reader recognizes the tag and it serves necessary information through speaker or bluetooth. And also it is possible for the blind to use less expensive one with RFID tag than the white stick with RFID Reader.