• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.14 no.8
• /
• pp.1819-1826
• /
• 2010

This paper proposes the image segmentation algorithm to extracts a fish object from a fish image for fish image retrieval. The conventional algorithm using gray level similarity causes wrong image segmentation result in the boundary area of the object and the background with similar gray level. The proposed algorithm uses the reinforced edge and the adaptive block-based threshold for the boundary area with weak contrast and the virtual object to improve the eroded or disconnected object in the boundary area without contrast. The simulation results show that the percentage of extracting the visual-fine object from the test images is under 90% in the conventional algorithm while it is 97.7% in the proposed algorithms.

• Korean Journal of Remote Sensing
• /
• v.32 no.5
• /
• pp.471-482
• /
• 2016

This paper presents an extraction method for water body which uses block-based image partitioning and extension of water body boundaries to improve the performance of supervised classification for water body extraction. The Mahalanobis distance image is created by computing the spectral information of Normalized Difference Water Index (NDWI) and Near Infrared (NIR) band images over a training site within the water body in order to extract an initial water body area. To reduce the effect of noise contained in the Mahalanobis distance image, we apply mean curvature diffusion to the image, which controls diffusion coefficients based on connectivity strength between adjacent pixels and then extract the initial water body area. After partitioning the extracted water body image into the non-overlapping blocks of same size, we update the water body area using the information of water body belonging to water body boundaries. The update is performed repeatedly under the condition that the statistical distance between water body area belonging to water body boundaries and the training site is not greater than a threshold value. The accuracy assessment of the proposed algorithm was tested using KOMPSAT-2 images for the various block sizes between $11{\times}11$ and $19{\times}19$. The overall accuracy and Kappa coefficient of the algorithm varied from 99.47% to 99.53% and from 95.07% to 95.80%, respectively.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.4
• /
• pp.533-540
• /
• 2022

Specialists diagnose skin cancer using a dermatoscopy to detect skin cancer as early as possible, but it is difficult to determine accurate skin lesions because skin lesions have various shapes. Recently, the skin lesion segmentation method using deep learning, which has shown high performance, has a problem in segmenting skin lesions because the boundary between healthy skin and skin lesions is not clear. To solve these issues, the proposed method constructs a transformer block to effectively segment the skin lesion, and constructs an edge decoder for each layer of the network to segment the skin lesion in detail. Experiment results have shown that the proposed method achieves a performance improvement of 0.041 ~ 0.071 for Dic Coefficient and 0.062 ~ 0.112 for Jaccard Index, compared with the previous method.

• Journal of KIISE:Software and Applications
• /
• v.27 no.4
• /
• pp.441-447
• /
• 2000

Automatic word spacing is needed to solve the automatic indexing problem of the non-spaced documents and the space-insertion problem of the character recognition system at the end of a line. We propose a word spacing algorithm that automatically finds out word spacing positions. It is based on the recognition of Eojeol components by using the sentence partition and bidirectional longest-match algorithm. The sentence partition utilizes an extraction of Eojeol-block where the Eojeol boundary is relatively clear, and a Korean morphological analyzer is applied bidirectionally to the recognition of Eojeol components. We tested the algorithm on two sentence groups of about 4,500 Eojeols. The space-level recall ratio was 97.3% and the Eojeol-level recall ratio was 93.2%.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2012.07a
• /
• pp.124-126
• /
• 2012

3차원 동영상은 수많은 응용분야에서 차세대 멀티미디어 콘텐츠로 큰 주목을 받고 있지만 2차원에서 3차원 콘텐츠로의 과도기인 현 시점에서 3차원 콘텐츠의 부족현상은 관련 산업분야의 큰 문제점으로 떠오르고 있다. 2D-to-3D 변환 기술은 높은 비용 없이 기존의 방대한 2차원 콘텐츠를 3차원 콘텐츠로 재사용하는 것이 가능하기 때문에 3차원 콘텐츠의 부족 문제를 해결할 수 있는 기술로 큰 관심을 끌고 있다. 본 논문에서는 블록 단위 초점 정보의 보간을 통한 새로운 초점 기반 3차원 변환 기법을 제안한다. 기존의 화소 단위 초점 측정치를 그대로 깊이 정보로 이용하는 경우나 분할 영역 단위 초점 측정치를 깊이 정보로 활용하는 경우는 이웃 화소간 깊이정보의 불연속성이 발생할 수 있지만 제안방식은 블록 초점 측청치의 보간수행으로 인해 이웃 화소 사이에서 뿐만 아니라 블록경계 혹은 영역경계에서도 특별할 스무딩필터 처리 없이도 화소간 깊이정보의 불연속성이 발생하지 않게 된다. 실험결과를 통해 제안한 방식이 기존의 방식들 보다 더 자연스러운 깊이 정보 추출 성능을 보여준다는 것을 알 수 있다.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.4 no.3
• /
• pp.126-133
• /
• 2003

This paper is related to a morphological segmentation method for extracting the moving object in video sequence using global motion compensation and two-dimensional spatio-temporal entropic thresholding. First, global motion compensation is performed with camera panning vector estimated in the hierarchical pyramid structure constructed by wavelet transform. Secondly, the regions with high possibility to include the moving object between two consecutive frames are extracted block by block from the global motion compensated image using two-dimensional spatio-temporal entropic thresholding. Afterwards, the LUT classifying each block into one among changed block, uncertain block, stationary block according to the results classified by two-dimensional spatio-temporal entropic thresholding is made out. Next, by adaptively selecting the initial search layer and the search range referring to the LUT, the proposed HBMA can effectively carry out fast motion estimation and extract object-included region in the hierarchical pyramid structure. Finally, after we define the thresholded gradient image in the object-included region, and apply the morphological segmentation method to the object-included region pixel by pixel and extract the moving object included in video sequence. As shown in the results of computer simulation, the proposed method provides relatively good segmentation results for moving object and specially comes up with reasonable segmentation results in the edge areas with lower contrast.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.10d
• /
• pp.550-552
• /
• 2002

본 논문에서는 웨이브렛 변환 영역의 정보를 신경망 학습을 통하여 영상검색에 관한 연구를 제안하였다. 영상검색은 연구가 이루어지고 있지만, 영상의 특징을 정확하게 표현한다는 것은 현실적으로 어렵기 때문에 영상의 저장 및 검색에 많은 어려움이 있다. 따라서 영상데이터의 효율적인 저장 및 검색을 위해서는 공간 영역보다는 변환 영역에서의 특징추출 방법이 고려되어야 한다. 본 논문에서는 웨이브렛 변환 후 생성되는 저주파 대역의 영상을 일정한 크기로 ( 2$^n$$\times$2$^n$) 분할한 다음 각 블록의 표준편차를 구하고, 주어진 경계 값을 기준으로 작성된 블록 맵을 유사성의 척도로 이용하여 유사한 영상을 함께 모아 카테고리 분류에 의한 저장을 한다. 또한 질의영상에 대한 블록 맵을 신경망 학습을 통해 해당 카테고리를 찾아 1:1매칭을 통한 검색을 함으로써 검색 시간을 줄이고, 제안된 시스템 효율을 증대 시킬 수 있었다.

• Annual Conference of KIPS
• /
• 2002.11a
• /
• pp.57-60
• /
• 2002

본 논문에서는 능동윤곽모델(active contour model)의 잘 알려져 있는 스네이크(snake) 알고리즘을 스테레오영상에 적용하여 좌 우 영상의 disparity 정보를 이용 객체의 경계선을 찾는 알고리즘을 제안한다. 스네이크는 객체의 경계를 얻기 위해 에지정보를 사용하는데 실제 이미지에서 객체의 경계가 아닌 인접한 주위의 강한 애지(edge)에 대해서도 영향을 받게 되는 문제가 있다. 이러한 문제를 해결하기 위해 스테레오영상의 disparity 정보를 이용하여 이를 개선하고 disparity 측정에 사용되는 블록매칭(block matching)방법을 스네이크 알고리즘에 적용시켰다.

• Journal of IKEEE
• /
• v.13 no.4
• /
• pp.1-7
• /
• 2009

A video segmentation algorithm, which can extract objects even with non-stationary backgrounds, is proposed in this work. The proposed algorithm is composed of three steps. First, we perform an initial segmentation interactively to build the probability density functions of colors per each macro block via kernel density estimation. Then, for each subsequent frame, we construct a coherence strip, which is likely to contain the object contour, by exploiting spatio-temporal correlations. Finally, we perform the segmentation by minimizing an energy function composed of color, coherence, and smoothness terms. Experimental results on various test sequences show that the proposed algorithm provides accurate segmentation results.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.8 no.8
• /
• pp.1740-1747
• /
• 2004

In this paper, we proposed an effective method for detecting fingerprint ridge direction based on the run-length and chain codes. First, a fingerprint image is normalized, and it is thresholded to obtain binary image with foreground and background regions. The foreground regions is composed of fingerprint ridges, and the ridges is encoded with the run-length and chain codes. To detect directional information, the boundary of ridge codes is traced, and curvature is calculated at ecah point of boundary. And the detected direction value is smoothed with appropriate window locally. The proposed method is applied to NIST and FVC2002 fingerprint database to evaluate performance. By the experimental results, the proposed method can be used to obtain ridge direction value in fingerprint image.