• Journal of the Korean Institute of Intelligent Systems
• /
• v.25 no.1
• /
• pp.91-96
• /
• 2015

In this study, we introduce robust face recognition system with illumination variation realized with the aid of CT preprocessing method. As preprocessing algorithm, Census Transform(CT) algorithm is used to extract locally facial features under unilluminated condition. The dimension reduction of the preprocessed data is carried out by using $(2D)^2$PCA which is the extended type of PCA. Feature data extracted through dimension algorithm is used as the inputs of proposed radial basis function neural networks. The hidden layer of the radial basis function neural networks(RBFNN) is built up by fuzzy c-means(FCM) clustering algorithm and the connection weights of the networks are described as the coefficients of linear polynomial function. The essential design parameters (including the number of inputs and fuzzification coefficient) of the proposed networks are optimized by means of artificial bee colony(ABC) algorithm. This study is experimented with both Yale Face database B and CMU PIE database to evaluate the performance of the proposed system.

• Journal of KIISE:Software and Applications
• /
• v.37 no.1
• /
• pp.9-18
• /
• 2010

This paper proposes a deformable registration method using a demon algorithm for aligning the lungs between end-exhale and end-inhale CT scans. The lungs are globally aligned by affine transformation and locally deformed by a demon algorithm. The use of floating gradient force allows a fast convergence in the lung regions with a weak gradient of the reference image. The active-cell-based demon algorithm helps to accelerate the registration process and reduce the probability of deformation folding because it avoids unnecessary computation of the displacement for well-matched lung regions. The performance of the proposed method was evaluated through comparisons of methods that use a reference gradient force or a combined gradient force as well as methods with and without active cells. The results show that the proposed method can accurately register lungs with large deformations and can reduce the processing time considerably.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2007.11a
• /
• pp.520-529
• /
• 2007

간암은 세계적으로 흔한 악성 종양에 속하지만 우리나라에서 간암은 위암, 폐암 다음으로 높은 사망률을 보이며 이러한 간암은 조기진단이 요구된다. 전문의는 간암의 진단을 위해 조영증강 CT영상을 이용하여 육안으로 간암을 판별하는데, 조영증강 CT영상을 이용한 진단은 주 종양의 진단에는 도움이 되지만 주 종양에서 주위 간 조직으로 전이된 간암들을 판별하는 것은 어려우며 실제로 시술 중에야 전이된 간암의 존재를 알 수 있다. 본 논문에서는 조영증강 CT영상을 이용하여 간과 주 종양을 자동으로 추출한 후, 미세하게 주 종양 주위로 전위된 간암들을 추출하는 방법을 제안하여 전문의를 보조할 수 있는 보조 전문가 시스템으로서의 유용성을 확인하고자한다. 조영증강 CT영상은 흉부에서 5mm간격으로 40 ${\sim}$ 50장정도로 촬영된다. 조영증강 CT영상을 이용하여 간 영역을 추출하기 위해서 간의 형태학적 정보 그리고 명암도와 명암의 분포도를 이용한 양자화 기법 등을 적용하여 추출하며 추출된 간 영역에서 간암의 후보 영역 추출은 간암의 명암도와 형태학적 특징 정보를 이용하여 추출한다. 본 논문에서는 간암의 추출을 위해 맵 상에 흩어져 분포되어 있는 유사 패턴들의 무게 중심을 찾아 하나의 패턴으로 그룹화 하는 개선된 SOM 알고리즘을 제안하여 간암 판별에 적용한 후, 기존의 SOM 알고리즘과 비교 분석한 결과. 본 논문에서 제안된 SOM 알고리즘을 적용한 간암 추출이 더 효율적임을 확인 할 수 있었으며, 전문의가 판별한 것과 비교 분석한 결과, 전문의를 보조할 수 있는 보조 전문가 시스템으로서의 가능성을 확인할 수 있었다.

• Journal of the Korean Society of Radiology
• /
• v.16 no.2
• /
• pp.131-139
• /
• 2022

The purpose of this study was to insert a pedicle screw into a pig thoracic vertebrae, a general CT scan(Non MAR), and a thoracic axial image obtained with the Metallic Artifact Reduction for Orthopedic Implants (O-MAR) to reduce artifacts. The image obtained by reconstructing the algorithm (Standard, Soft, Bone, Detail) was used using the image J program. Signal to noise ratio(SNR) and contrast to noise ratio(CNR) were compared and analyzed by obtaining measured values based on the given equation. And this study was to investigate tube voltage and algorithm suitable for CT scan for thoracic pedicle screw insertion. As a result, when non-MAR was used, the soft algorithm showed the highest SNR and CNR at 80, 100, 120, and 140 kVp, On the other hand, when MAR was used, the standard algorithm showed the highest at 80 kVp, and the standard and soft algorithms showed similar values at 100 kVp. At 120 kVp, the Soft and Standard algorithms showed similar values, and at 140 kVp, the Soft algorithm showed the highest SNR and CNR. Therefore, when comparing Non-MAR and MAR, even if MAR was used, SNR and CNR did not increase in all algorithms according to the change in tube voltage. In conclusion, it is judged that it is advantageous to use the Soft algorithm at 80, 100, 120, and 140 kVp in Non MAR, the Standard algorithm at 80 and 100 kVp in MAR, and the Soft algorithm at 120 and 140 kVp. This study is expected to serve as an opportunity to further improve the quality of images by using selective tube voltage and algorithms as basic data to help evaluate images of pedicle screw CT scans in the future.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.05a
• /
• pp.312-315
• /
• 2022

CT의 반복재구성방법은 투영과 역투역을 번갈아가며 최적의 단면영상을 얻을 때까지 반복한다. 영상재구성 시간을 단축하기 위하여 시간이 많이 소요되는 투영을 빠르게 수행할 수 있는 알고리즘이 필요하다. 본 논문은 Siddon 알고리즘을 개선한 Jacobs 버전보다 대략 10% 빠른 알고리즘을 제안한다. 평행빔의 경우에 대해 조사되었지만 향후 부채살빔 및 콘빔의 경우로 확장이 가능하다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.10d
• /
• pp.526-528
• /
• 2002

본 논문은 CT 영상을 이용하여 체지방과 근육의 체적 및 비율을 분석하는 알고리즘 개발에 대하여 기술한다. CT 영상에는 체지방, 근육, 공기, 뼈등의 구성성분들이 서로 다른 명암값을 가지고 분포한다. 이 논문에서는 히스토그램을 통하여 각 구성성분에 대한 명암값을 찾아내었다. 찾아낸 명암값에 따라 체지방과 근육 그리고 뼈, 공기에 각각 색을 입혀 시각적으로 표현하여 체지방과 근육의 분포와 비율을 볼 수 있는 환경을 만들고 수치적으로 체적 및 비율의 결과 값을 출력하는 알고리즘을 개발하였다. 또한 단계적인 체지방 측정 프로그램을 위해 DB 환경을 구축하여 모든 자료를 저장하고 불러올 수 있는 환경을 만들어 체지방 측정 알고리즘을 완성하였다. 알고리즘은 돼지 실험에서 측정된 체지방과 비교하여 정확성을 검증한 결과 약 92.86%의 정확도를 보였다.

• Journal of KIISE:Software and Applications
• /
• v.28 no.2
• /
• pp.124-131
• /
• 2001

복부 CT 영상에서 위의 자동적인 추출에 대하여 연구하였다. 복부 CT 영상에서 여러 장기가 비슷한 명암 값을 나타내며 분포 해 있다. 본 논문에서는 복부 CT 영상의 여러 장기 가운데 위를 자동적으로 추출하는 알고리즘을 개발하였다. 위는 움직이는 장기이며, 음식물로 채워진 부분과 공기로 채원진 부분으로 나뉘어져 있다. 이를 바탕으로 히스토그램 분석을 통한 명암 값 정보와 위치 정보를 이용하여 위를 탐색하고, 주변 다른 장기를 제거하는 다듬기 과정으로 완전한 위 추출 알고리즘을 완성하였다. 또한 돼지 실험에서 추출된 위의 체적을 비교하여, 개발된 알고리즘의 정확성을 검증한 결과 약 95%의 정확도를 보였다.

• Journal of the Korean Society of Radiology
• /
• v.13 no.3
• /
• pp.481-487
• /
• 2019

Based on the scan conditions and algorithms that are generally applied during examinations during head CT examinations, the results of dose reduction through the application of algorithm changes were investigated through experiments. As a result, the dose reduction effect was more meaningful for the change of perfusion than for the tube voltage, and the quality evaluation using the brain phantom was relatively less reduced when the dose was reduced after the application of the Bone algorithm, especially for the application of the Bone algorithm, and the deviation of the mean CT number or Pixel value was measured relatively significantly. In other words, the conditions under which dose was reduced and quality was maintained to reduce the patient's exposure dose and obtain images of the same quality were obtained with the application of the Smooth algorithm and the resulting values of 120 kVp, 160 mA. At this point, doses were reduced by about 28%, and the mean CT number or Pixel value was also measured with relatively little error. If the results are applied to patients who visit the hospital for examination or follow-up after applying various algorithms and follow up scan conditions, the results are considered to be very useful in reducing patient exposure dose.

• Proceedings of the Korean Society of Medical Physics Conference
• /
• 2004.11a
• /
• pp.104-107
• /
• 2004

In this study, using brain phantom for multi-modality imaging, we acquired CT, MR and PET images and performed registration of these anatomical images and nuclear medicine functional images. The algorithms and program applied for registration were Chamfer Matching and Mutual Information Maximization algorithm which have been using frequently in clinic and verified accuracy respectively. In result, both algorithms were useful methods for CT-MR, CT-PET and MR-PET. But Mutual Information Maximization was more effective algorithm for low resolution image as nuclear medicine functional image.

• Journal of Radiation Protection and Research
• /
• v.36 no.2
• /
• pp.79-85
• /
• 2011

OThe purpose of this article was to measure and compare the value of the metal artifact reduction (MAR) algorithm by Dual energy(kVp switching) CT (Computed Tomography) for non using MAR and we introduced new variable Dual energy CT applications through a clinical scan. The used equipment was GE Discovery 750HD with Dual-Energy system(kVp switching). CT scan was performed on the neck and abdomen area subject for patients. Studies were from Dec 20 2010 to Feb 10 2011 and included 25 subject patients with prosthesis. We were measured the HU (Hounsfield Unit) and noise value at metal artifact appear(focal loss of signal and white streak artifact area) according to the using MAR algorithm. Statistical analyses were performed using the paired sample t-test. In patient subject case, the statistical difference of showing HU was p=0.01 and p=0.04 respectively. At maximum black hole artifact area and white streak artifact area according to the using MAR algorithm. However noise was p=0.05 and p=0.04 respectively; and not the affected black hole and white streak artifact area. Dual Energy CT with the MAR algorithm technique is useful reduce metal artifacts and could improve the diagnostic value in the diagnostic image evaluation of metallic implants area.