• Proceedings of the Korea Multimedia Society Conference
• /
• 2001.11a
• /
• pp.165-169
• /
• 2001

본 논문에서는 텍스쳐가 포함된 칼라 영상으로부터 텍스쳐에 무관하게 영역을 분할할 수 있는 방법을 개발하였다. 빠른 처리를 위해 영상을 블록 단위로 쪼개고 블록의 경계 성분값(H)을 계산하여 영역 분할에 이용할 수 있도록 하였다. M값은 객체의 경계에서는 높은 경계 강도를 갖지만 영역 내부나 텍스쳐 경계에서는 상대적으로 낮은 경계 강도를 갖도록 정의되었다 영상 분할을 위해 M값으로 표현된 M영상으로부터 Watershed를 이용해 경계 위치를 결정하고 닫혀진 형태로 경계가 표현될 수 있도록 하였다. 그런데 Watershed 방법은 과잉 분할 결과를 초래하므로 인접 영역 사이의 공유 경계에 대한 강도와 영역 내부의 칼라 분포 특성을 이용하여 영역을 병합함으로써 객체 경계처럼 중요한 변화가 발생되는 영역 단위의 최종 영상 분할된 결과를 얻을 수 있도록 하였다. 본 논문에서 제안한 방법은 MPEG4나 내용기반검색을 위한 영역분할에 유용하게 적용될 수 있을 것이다.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.58 no.1
• /
• pp.91-97
• /
• 2016

The objective of this study was to classify between radish and Chinese cabbage in autumn using hyperspectral images. The hyperspectral images were acquired by Compact Airborne Spectrographic Imager (CASI) with 1m spatial resolution and 48 bands covering the visible and near infrared portions of the solar spectrum from 370 to 1044 nm with a bandwidth of 14 nm. An object-based technique is used for classification of radish and Chinese cabbage. It was found that the optimum parameter values for image segmentation were scale 400, shape 0.1, color 0.9, compactness 0.5 and smoothness 0.5. As a result, the overall accuracy of classification was 90.7 % and the kappa coefficient was 0.71. The hyperspectral images can be used to classify other crops with higher accuracy than radish and Chines cabbage because of their similar characteristic and growth time.

• Journal of Korea Multimedia Society
• /
• v.18 no.10
• /
• pp.1149-1156
• /
• 2015

This paper proposes the hand contour detector which is robust to noises. Existing methods reduce noises by applying morphology to extracted edges, detect finger tips by using the center of hands, or exploit the intersection of curves from hand area candidates based on J-value segmentation(JSEG). However, these approaches are so vulnerable to noises that are prone to detect non-hand parts. We propose the noise tolerant hand contour detection method in which non-skin area noises are removed by applying skin area detection, contour detection, and a threshold value. By using the implemented system, we observed that the system was successfully able to detect hand contours.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2002.05d
• /
• pp.553-558
• /
• 2002

본 논문에서는 컬러 영상에서 피부색 분할과 소벨 마스크를 이용한 얼굴 영역 검출 알고리즘을 제안한다. 제안된 알고리즘은 YCbCr색공간에서 Cb와 Cr성분을 이용하여 피부색 분할을 한 후에 형태학적 필터링과 레이블링을 통해 얼굴 후보 영역을 분리한다. 분리된 각 후보 영역에 대해 휘도 성분 Y에서 소벨 마스크의 수직 연산자를 적용한 후에 수평 투영을 통해 나타난 최대값을 눈의 위치로 검출해낸다. 비슷하게 얼굴의 지형적인 특징과 소벨 마스크의 수평 연산자를 적용하여 계산된 수평 투영의 최대값에 따라 턱 부분을 검출한다. 컴퓨터 시뮬레이션 결과는 제안된 방법이 기존의 방법보다 얼굴 영역을 정확하게 분리할 수 있음을 보인다.

• Proceedings of the KIEE Conference
• /
• 2005.07d
• /
• pp.2780-2782
• /
• 2005

In this paper, we present a mobile robot localization solution by using a Ubiquitous Vision System (UVS). The collective information gathered by multiple cameras that are strategically placed has many advantages. For example, aggregation of information from multiple viewpoints reduces the uncertainty about the robots' positions. We construct UVS as a multi-agent system by regarding each vision sensor as one vision agent (VA). Each VA performs target segmentation by color and motion information as well as visual tracking for multiple objects. Our modified identified contractnet (ICN) protocol is used for communication between VAs to coordinate multitask. This protocol raises scalability and modularity of thesystem because of independent number of VAs and needless calibration. Furthermore, the handover between VAs by using ICN is seamless. Experimental results show the robustness of the solution with respect to a widespread area. The performance in indoor environments shows the feasibility of the proposed solution in real-time.

• 한국HCI학회:학술대회논문집
• /
• 2007.02a
• /
• pp.630-635
• /
• 2007

In this paper, we propose a system for multiple people tracking using fragment based histogram matching. Appearance model is based on IHLS color histogram which can be calculated efficiently using integral histogram representation. Since histograms will loss all spatial information, we define a fragment based region representation which retain spatial information, robust against occlusion and scale issue by using disparity information. Multiple people labeling is maintained by creating online appearance representation for each people detected in scene and calculating fragment vote map. Initialization is performed automatically from background segmentation step.

• Proceedings of the IEEK Conference
• /
• 2009.05a
• /
• pp.34-36
• /
• 2009

This paper proposes an effective lane detection algorithm using hugh transform and lane model. The proposed lane detection algorithm includes two major components, i.e., lane marks segmentation and an exact lane extraction using a novel postprocessing technique. The first step is to segment lane marks from background images using HSV color model. Then, a novel postprocessing is used to detect an exact lane using Hugh transform and lane models(linear and curved lane models). The postprocessing consists of three parts, i.e, thinning process, Hugh Transform and filtering process. We divide input image into three regions of interests(ROIs). Based on lane curve function(LCF), we can detect an exact lane from various extracted lane lines. The lane models(linear and curved lane mode]) are used in order to judge whether each lane segment is fit or not in each ROIs. Experimental results show that the proposed scheme is very effective in lane detection.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.105.2-105
• /
• 2001

This paper studies continuous Korean Sign Language (KSL) recognition using color vision. In recognizing gesture words such as sign language, it is a very difficult to segment a continuous sign into individual sign words since the patterns are very complicated and diverse. To solve this problem, we disassemble the KSL into 18 hand motion classes according to their patterns and represent the sign words as some combination of hand motions. Observing the speed and the change of speed of hand motion and using state automata, we reject unintentional gesture motions such as preparatory motion and meaningless movement between sign words. To recognize 18 hand motion classes we adopt Hidden Markov Model (HMM). Using these methods, we recognize 5 KSL sentences and obtain 94% recognition ratio.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.8 no.1
• /
• pp.55-64
• /
• 2012

The advantage of direct volume rendering is to visualize structures of interest in the volumetric data. However it is still difficult to simultaneously show interior and exterior structures. Recently, cone beam computed tomography(CBCT) has been used for dental diagnosis. Despite of its usefulness, there is a limitation in the detection of interior structures such as pulp and inferior alveolar nerve canal. In this paper, we propose an efficient volume rendering model for visualizing important interior as well as exterior structures of dental CBCT. It is based on the concept of illustrative volume rendering and enhances boundary and silhouette of structures. Moreover, we present a new method that assigns a different color to structures in the rear so as to distinguish the front ones from the rear ones. This proposed rendering model has been implemented on graphics hardware, so that we can achieve interactive performance. In addition, we can render teeth, pulp and canal without cumbersome segmentation step.

• International Journal of Contents
• /
• v.8 no.1
• /
• pp.30-38
• /
• 2012

Text extraction and binarization are the important pre-processing steps for text recognition. The performance of text binarization strongly related to the accuracy of recognition stage. In our proposed method, the first stage based on line detection and shape feature analysis applied to locate the position of a business card and detect the shape from the complex environment. In the second stage, several local regions contained the possible text components are separated based on the projection histogram. In each local region, the pixels grouped into several connected components based on the connected component labeling and projection histogram. Then, classify each connect component into text region and reject the non-text region based on the feature information analysis such as size of connected component and stroke width estimation.