• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2002년도 ITC-CSCC -1
• /
• pp.670-673
• /
• 2002

This paper proposes a digital algorithm for gray-scale/color image segmentation of real-time video signals and a cell-network-based implementation architecture in state-of-the-art CMOS technology. Through extrapolation of design and simulation results we predict that about 300$\times$300 pixels can be integrated on a chip at 100nm CMOS technology, realizing very high-speed segmentation at about 1600sec per color image. Consequently real-time color-video segmentation will become possible in near future.

• 한국멀티미디어학회논문지
• /
• 제4권4호
• /
• pp.234-332
• /
• 2001

프랙탈 영상 부호화의 새로운 응용 분야인 프랙탈 영역 분할법의 YST방법은 주기점에 의한 라벨 붙이기와 프랙탈 변환에 의한 라벨 수정을 병용한 영역 분할법을 제안하였다. 그러나 이 개선법은 영역 분할의 질적인 개선은 가능하였으나, 여전히 라벨 붙이기와 라벨 수정의 과정에서 중복성이 남아 있다. 이 문제점의 해결방안으로 본 논문에서는 궤도에 따른 라벨 붙이기와 프랙탈 변환의 반복 과정에 관한 제약 조건을 제안한다.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 1997년도 추계학술대회 논문집
• /
• pp.129-132
• /
• 1997

Laser scanning is widely used due to its fast measuring and high precision, and the segmentation of the scanned data is necessary for the fast and efficient surface modelling. But most segmentation techniques are based on the very regular data and the adaptation of previous techniques to the scanned data does not usually produce good result. A new approach to perform the segmentation on the scanned data is introduced to deal with problems during reverse engineering process. The approach is based on the triangulated data and its result is depending on the some user-defined criteria. The result is illustrated to demonstrate its adaptability to the measured data on free-form surface and the each result by different criteria is compared respectively.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송공학회 2009년도 IWAIT
• /
• pp.148-153
• /
• 2009

Few methods have dealt with segmenting multiple images with analogous content. Concurrent images of a scene and gathered images of a similar foreground are examples of these images, which we term consistent scene images. In this paper, we present a method to segment these images based on manual segmentation of one image, by iteratively propagating information via multi-level cues with adaptive confidence. The cues are classified as low-, mid-, and high- levels based on whether they pertain to pixels, patches, and shapes. Propagated cues are used to compute potentials in an MRF framework, and segmentation is done by energy minimization. Through this process, the proposed method attempts to maximize the amount of extracted information and maximize the consistency of segmentation. We demonstrate the effectiveness of the proposed method on several sets of consistent scene images and provide a comparison with results based only on mid-level cues [1].

• 한국군사과학기술학회지
• /
• 제25권6호
• /
• pp.606-616
• /
• 2022

Semantic segmentation in autonomous driving for unstructured environments is challenging due to the presence of uneven terrains, unstructured class boundaries, irregular features and strong textures. Current off-road datasets exhibit difficulties like class imbalance and understanding of varying environmental topography. To overcome these issues, we propose a deep learning framework for semantic segmentation that involves a pooled class semantic segmentation with five classes. The evaluation of the framework is carried out on two off-road driving datasets, RUGD and TAS500. The results show that our proposed method achieves high accuracy and real-time performance.

• Investigative Magnetic Resonance Imaging
• /
• 제19권2호
• /
• pp.88-98
• /
• 2015

Purpose: To compare the interobserver and intraobserver reliability of mean apparent diffusion coefficient (ADC) values using contrast-enhanced (CE) T1 weighted image (WI) and T2WI as structural images between manual and semiautomatic segmentation methods. Materials and Methods: Between January 2011 and May 2013, 28 patients who underwent brain MR with diffusion weighted image (DWI) and were pathologically confirmed as having glioblastoma participated in our study. The ADC values were measured twice in manual and semiautomatic segmentation methods using CE-T1WI and T2WI as structural images to obtain interobserver and intraobserver reliability. Moreover, intraobserver reliabilities of the different segmentation methods were assessed after subgrouping of the patients based on the MR findings. Results: Interobserver and intraobserver reliabilities were high in both manual and semiautomatic segmentation methods on CE-T1WI-based evaluation, while interobserver reliability on T2WI-based evaluation was not high enough to be used in a clinical context. The intraobserver reliability was particularly lower with the T2WI-based semiautomatic segmentation method in the subgroups with involved $lobes{\leq}2$, with partially demarcated tumor borders, poorly demarcated inner margins of the necrotic portion, and with perilesional edema. Conclusion: Both the manual and semiautomatic segmentation methods on CE-T1WI-based evaluation were clinically acceptable in the measurement of mean ADC values with high interobserver and intraobserver reliabilities.

• Journal of information and communication convergence engineering
• /
• 제13권3호
• /
• pp.205-214
• /
• 2015

In this paper, we propose a new approach to learning a discriminative model for figure/ground segmentation by incorporating the bag-of-features and conditional random field (CRF) techniques. We advocate the use of image patches instead of superpixels as the basic processing unit. The latter has a homogeneous appearance and adheres to object boundaries, while an image patch often contains more discriminative information (e.g., local image structure) to distinguish its categories. We use pixel-level sparse coding to represent an image patch. With the proposed feature representation, the unary classifier achieves a considerable binary segmentation performance. Further, we integrate unary and pairwise potentials into the CRF model to refine the segmentation results. The pairwise potentials include color and texture potentials with neighborhood interactions, and an edge potential. High segmentation accuracy is demonstrated on three benchmark datasets: the Weizmann horse dataset, the VOC2006 cow dataset, and the MSRC multiclass dataset. Extensive experiments show that the proposed approach performs favorably against the state-of-the-art approaches.

• 대한전자공학회논문지TE
• /
• 제37권3호
• /
• pp.45-52
• /
• 2000

본 연구에서는 인체 상반신영상에서 얼굴부위를 분할하기 위한 영상분할 알고리즘을 제안하였다. 제안한 알고리즘은 HWT를 적용하여 영상의 경계를 이루는 차분영상인 고주파대역과 평균영상인 저주파대역으로 분리하고, 저주파대역에서 고립점과 돌출부위, 경계중복점을 제거하였다. 또한 제안한 경계검출 알고리즘으로 경계를 검출하고 단순화시켰으며, 1픽셀 단위의 세선화과정을 통하여 경계를 선명하게 하였다. 그리고 제안 한 폐곡선추적 알고리즘으로 얼굴부위 경계만을 추출한 뒤, 마스크를 구성하고 원영상과의 정합을 통하여 얼굴부위분할을 하였다. 제안한 알고리즘을 적용하여 얼굴부위 분할실험을 실행한 결과 95.88%의 분할값을 갖는 얼굴분할이 이루어졌다.

• 한국의류학회지
• /
• 제35권12호
• /
• pp.1440-1453
• /
• 2011

Proper consumer segmentation is receiving more attention from industry professionals as markets become more diverse and consumer-centered. Researchers have recognized the limitations of the traditional cluster analysis technique and this research study analyzes market segmentation using Mixture-model or latent-class segmentation. This study used a questionnaire to determine the characteristics of clothing shoppers using a new technique that proved its superiority over traditional techniques. Questions included items measuring fashion shopping behavior, store choice criteria, apparel consumption styles, price perception by product type, and demographic characteristics. Data were collected from 1074 males and females in their 20s and 30s through an online survey. SPSS 16.0 and Latent GOLD 4.0 were used to analyze the data. The ideal typology of clothing shoppers using the Mixture-model were: 'brand loyalty orientated group', 'group of conservative late 30s', 'group of pleasure-emotion early 20s', 'value oriented consumer product with high-income group', 'group of eco/symbol oriented consumer', and 'group of utility/goal oriented male consumer'. This study showed differences in fashion product purchasing behavior by conducting market segmentation for clothing shoppers using the Mixture-model.

• 대한음성학회지:말소리
• /
• 제58호
• /
• pp.67-81
• /
• 2006

For a large corpus of time-aligned data, HMM based approaches are most widely used for automatic segmentation, providing a consistent and accurate phone labeling scheme. There are two methods for training in HMM. Flat starting method has a property that human interference is minimized but it has low accuracy. Bootstrap method has a high accuracy, but it has a defect that manual segmentation is required In this paper, a new algorithm is proposed to minimize manual work and to improve the performance of automatic segmentation. At first phase, voiced, unvoiced and silence classification is performed for each speech data frame. At second phase, the phoneme sequence is aligned dynamically to the voiced/unvoiced/silence sequence according to the acoustic phonetic rules. Finally, using these segmented speech data as a bootstrap, phoneme model parameters based on HMM are trained. For the performance test, hand labeled ETRI speech DB was used. The experiment results showed that our algorithm achieved 10% improvement of segmentation accuracy within 20 ms tolerable error range. Especially for the unvoiced consonants, it showed 30% improvement.