• 한국정밀공학회지
• /
• 제16권5호통권98호
• /
• pp.48-57
• /
• 1999

Applications of thresholding technique are based on the assumption that object and background pixels in a digital image can be distinguished by their gray level values. For the segmentation of more complex images, it is necessary to resort to multiple threshold selection techniques. This paper describes a new method for multiple threshold selection of gray level images which are not clearly distinguishable from the background. The proposed method consists of three main stages. In the first stage, a probability distribution function for a gray level histogram of an image is derived. Cluster points are defined according to the probability distribution function. In the second stage, fuzzy partition matrix of the probability distribution function is generated through the fuzzy clustering process. Finally, elements of the fuzzy partition matrix are classified as clusters according to gray level values by using max-membership method. Boundary values of classified clusters are selected as multiple threshold. In order to verify the performance of the developed algorithm, automatic inspection process of ball grid array is presented.

• 대한의용생체공학회:의공학회지
• /
• 제20권2호
• /
• pp.149-154
• /
• 1999

자기공명영상은 다른 의료영상에 비해서 보다 정확한 해부학적인 진단 정보를 제공해 주므로 널리 이용되고 있다. 본 논문에서는 이차원 축단면 뇌 자기총명영상을 분할하는 자동화 알고리즘과 병별에 의해서 손상된 슬라이스를 검출하는 알고리즘을 제안하였다. 영상분활 과정은 두단계로 구성되어 있는데, 첫 단계에서는 이진화와 형태학적 연산을 이용하여 대뇌영역을 추출하고, 둘째 단계에서는 FCM(Fuzzy C-means)알고리즘을 이용하여 추출된 대뇌 내부의 각 조직을 분할하였다. FCM알고리즘은 분할하는 조직의 수가 증가할수록 급격하게 많은 실행시간을 요구하므로 제안하는 두단계 영상분할 과정을 통하여 실행시간을 향상시켰다. 병변 인식은 해부학적지식과 패턴매칭을 이용하였다.

• 대한의용생체공학회:의공학회지
• /
• 제18권4호
• /
• pp.357-364
• /
• 1997

유방암은 그 조기 발견이 암환자의 사망률을 줄이는 데 있어서 가장 중요한 요소임을 알려져 있다. 스크리닝 검사에 의해 발견되는 유방암의 20%정도를 차지하는 DCIS(ductal carcinoma in situ)의 경우 미세석회화만이 필름 상에서 볼 수 있는 유일한 소견이다. 따라서 미세석회화를 발견하고 그 형태와 분포의 분석을 통한 진단이 암의 조기 발견에 매우 중요하다. 이 검출과정을 자동화하려는 시도가 디지털 영상처리 기술의 관심이 되어 왔다. 본 연구에서는 상관계수를 특징(feature)으로 사용하여 성능을 향상시킨 통계적 패턴 분류법을 제안하였다. 결과적인 검출율은 통계적 문턱치 설정에 의한 이진호 방법과 비교하여 48%에서 83%로 향상되었다. 성능은 TP와 FP로 평가되었으며 클래스 구분시의 오차도 함께 나타내었다.

• 대한의용생체공학회:의공학회지
• /
• 제20권1호
• /
• pp.59-68
• /
• 1999

본 논문에서는 뇌량의 형태정보와 통계적 특성을 이용한 중앙시상 두뇌자기공명영상의 뇌량자동인식 알고리즘을 제안한다. 제안된 알고리즘에서는 우선 뇌량의 통계적 특성에 일치하는 영역들을 추출하고 형태정보와 일치하는 영역을 검출한다. 이러한 형태정합을 위해 기존의 윤곽정합알고리즘 대신에 통계적인 특성을 적응적으로 변화시켜 형태정보와 일치하는 영역을 검출하는 방향성 창영역확장 알고리즘을 제안하였다. 실험결과 제안된 알고리즘의 우수성을 확인할 수 있었다.

• 센서학회지
• /
• 제17권5호
• /
• pp.338-345
• /
• 2008

Steel bearing balls are important component in machines having moving parts. In this paper we describe a vision-based automatic inspection system designed for sensing defects on the surface of steel bearing balls. The system has a camera looking down over a rail on which balls roll. Two mirrors are installed at both sides of the rail so that the side parts of a ball can be well inspected. The entire ball surface can be sufficiently seen by taking three images at $120^{\circ}$ rotation interval. Defects are detected by thresholding the difference image between an image captured and the reference image of a good ball.

• 한국군사과학기술학회지
• /
• 제17권2호
• /
• pp.175-181
• /
• 2014

In this paper, we propose a robust clutter pre-thresholding method using background adaptive Threshold Map for the clutter rejection in the complex coastal environment. The proposed algorithm is composed of the use of Threshold Map's and method of its calculation. Additionally we also suggest an automatic decision method of Thresold Map's update. Experimental results on some sets of real infrared image sequence show that the proposed method could remove clutters effectively without any loss of detection rate for the aim target and reduce processing time dramatically.

• Journal of Information Processing Systems
• /
• 제13권5호
• /
• pp.1135-1148
• /
• 2017

Speech recognition is one of the fascinating fields in the area of Computer science. Accuracy of speech recognition system may reduce due to the presence of noise present in speech signal. Therefore noise removal is an essential step in Automatic Speech Recognition (ASR) system and this paper proposes a new technique called combined thresholding for noise removal. Feature extraction is process of converting acoustic signal into most valuable set of parameters. This paper also concentrates on improving Mel Frequency Cepstral Coefficients (MFCC) features by introducing Discrete Wavelet Packet Transform (DWPT) in the place of Discrete Fourier Transformation (DFT) block to provide an efficient signal analysis. The feature vector is varied in size, for choosing the correct length of feature vector Self Organizing Map (SOM) is used. As a single classifier does not provide enough accuracy, so this research proposes an Ensemble Support Vector Machine (ESVM) classifier where the fixed length feature vector from SOM is given as input, termed as ESVM_SOM. The experimental results showed that the proposed methods provide better results than the existing methods.

• 한국측량학회지
• /
• 제37권5호
• /
• pp.351-357
• /
• 2019

UAV (Unmanned Aerial Vehicle) images can be exploited for rapid forest fire damage assessment by virtue of UAV systems' advantages. In 2019, catastrophic forest fire occurred in Goseong and Sokcho, Korea and burned 1,757 hectares of forests. We visited the town in Goseong where suffered the most severe damage and conducted UAV flights for forest fire damage assessment. In this study, economic and rapid damage assessment method for forest fire has been proposed using UAV systems equipped with only a RGB sensor. First, forest masking was performed using automatic elevation thresholding to extract forest area. Then ExG (Excess Green) vegetation index which can be calculated without near-infrared band was adopted to extract damaged forests. In addition, entropy filtering was applied to ExG for better differentiation between damaged and non-damaged forest. We could confirm that the proposed forest masking can screen out non-forest land covers such as bare soil, agriculture lands, and artificial objects. In addition, entropy filtering enhanced the ExG homogeneity difference between damaged and non-damaged forests. The automatically detected damaged forests of the proposed method showed high accuracy of 87%.

• Korean Journal of Radiology
• /
• 제20권8호
• /
• pp.1275-1284
• /
• 2019

Objective: To develop algorithms using convolutional neural networks (CNNs) for automatic segmentation of acute ischemic lesions on diffusion-weighted imaging (DWI) and compare them with conventional algorithms, including a thresholding-based segmentation. Materials and Methods: Between September 2005 and August 2015, 429 patients presenting with acute cerebral ischemia (training:validation:test set = 246:89:94) were retrospectively enrolled in this study, which was performed under Institutional Review Board approval. Ground truth segmentations for acute ischemic lesions on DWI were manually drawn under the consensus of two expert radiologists. CNN algorithms were developed using two-dimensional U-Net with squeeze-and-excitation blocks (U-Net) and a DenseNet with squeeze-and-excitation blocks (DenseNet) with squeeze-and-excitation operations for automatic segmentation of acute ischemic lesions on DWI. The CNN algorithms were compared with conventional algorithms based on DWI and the apparent diffusion coefficient (ADC) signal intensity. The performances of the algorithms were assessed using the Dice index with 5-fold cross-validation. The Dice indices were analyzed according to infarct volumes (< 10 mL, ≥ 10 mL), number of infarcts (≤ 5, 6-10, ≥ 11), and b-value of 1000 (b1000) signal intensities (< 50, 50-100, > 100), time intervals to DWI, and DWI protocols. Results: The CNN algorithms were significantly superior to conventional algorithms (p < 0.001). Dice indices for the CNN algorithms were 0.85 for U-Net and DenseNet and 0.86 for an ensemble of U-Net and DenseNet, while the indices were 0.58 for ADC-b1000 and b1000-ADC and 0.52 for the commercial ADC algorithm. The Dice indices for small and large lesions, respectively, were 0.81 and 0.88 with U-Net, 0.80 and 0.88 with DenseNet, and 0.82 and 0.89 with the ensemble of U-Net and DenseNet. The CNN algorithms showed significant differences in Dice indices according to infarct volumes (p < 0.001). Conclusion: The CNN algorithm for automatic segmentation of acute ischemic lesions on DWI achieved Dice indices greater than or equal to 0.85 and showed superior performance to conventional algorithms.

• 대한방사선기술학회지:방사선기술과학
• /
• 제32권1호
• /
• pp.45-52
• /
• 2009

유방 촬영술(Digital mammography)은 유방암의 조기 진단에서 매우 중요한 진단 방법으로서 비촉지성 유방암의 조기 발견율을 높여 유방암에 따른 여성의 사망률을 감소시키고 있다. 그 중에서도 유방 병변의 미세석회화(Microcalcification)는 조기 유방암의 진단에 있어서 중요한 병변으로 보고 되고 있으며, 선별 검사로 임상적 유용성이 확립된 상태이다. 유방 촬영술에서 미세석회화 소견은 영상의학과 전문의가 판독하여 조직 검사에서 양성 및 악성 병변에 대하여 각각 군집의 개수, 군집 당 석회화 수, 미세석회화 크기와 범위, 미세석회화 형태, 동반 종괴의 유무 등을 분석하여 최종적으로 진단을 확정한다. 그러므로 군집화된 미세석회화의 정보는 유방암 예측에 있어 임상적인 실질 정보를 가지고 있으며, 의사에게 진단을 위한 검사의 기본적인 가이드라인을 제시한다. 따라서 본 연구에서는 유방 촬영술의 디지털 영상에 나타난 미세석회화의 정량적인 계산을 위해서 DoG filter, Adaptive thresholding, Expectation Maximization의 3단계를 제안한다. 제안한 알고리듬을 실험을 통하여 군집화 및 각 클러스터 내의 미세석회화의 분포 개수, 길이를 측정하였으며, 임상의 사에게 디지털 유방영상의 분석을 통하여 초기 유방암 진단의 지표를 제시할 것으로 사료된다. 그리고 이는 객관적인 유방암 컴퓨터자동검출(CAD)에 사용될 수 있는 병변의 정보로서 가능성을 보였다.