• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.11a
• /
• pp.903-906
• /
• 2021

인공지능 기술 발전으로, 의료영상 분야에서도 딥러닝 기반 질병 진단 연구가 활발히 진행되고 있다. 딥러닝 모델 개발 시, 학습 데이터 품질은 모델의 성능과 신뢰성에 매우 큰 영향을 미친다. 그러나 의료 분야의 경우 도메인 지식에 대한 진입 장벽이 높아 개발자가 학습에 사용되는 의료영상 데이터의 품질을 평가하기 어렵다. 이로 인해, 많은 의료영상 분야에서는 각 분야의 특성(질병의 종류, 관찰 아나토미 등)에 따른 영상 품질 평가 방법을 제시해왔다. 그러나 기존의 방법은 특정 질병에 초점이 맞춰져, 일반화된 품질 평가 기준을 제시하고 있지 않다. 따라서 본 논문에서는 대부분의 흉부 질환을 진단하기 위한 흉부 X선 영상의 품질을 평가할 수 있는 기준을 제안한다. 우선, 흉부 X선 영상을 대상으로 관찰된 영역인 심장, 횡격막, 견갑골, 폐 등을 분할하여, 3D 히스토그램을 기반으로 각 영역별 통계적인 정밀 품질 평가 기준을 제안한다. 본 연구에서는 JSRT, Chest 14의 오픈 데이터셋을 활용하여 적용 실험을 수행하였으며, 민감도는 97.6%, 특이도는 92.8%의 우수한 성능을 확인하였다.

• The Journal of Korean Society for Radiation Therapy
• /
• v.24 no.1
• /
• pp.11-14
• /
• 2012

Purpose: Anisotropic Analytical Algorithm (AAA) provides more accurate dose calculation regarding impact on scatter and tissue inhomogeneity in comparison to Pencil Beam Convolution (PBC) algorithm. This study tries to analyze the difference of dose distribution according to PBC algorithm and dose calculation algorithm of AAA on breast cancer tangential plan. Materials and Methods: Computerized medical care plan using Eclipse treatment planning system (version 8.9, VARIAN, USA) has been established for the 10 breast cancer patients using 6 MV energy of Linac (CL-6EX, VARIAN, USA). After treatment plan of Conventional Radiation Therapy plan (Conventional plan) and Field-in-Field Intensity Modulated Radiation Therapy plan (FiF plan) using PBC algorithm has been established, MU has been fixed, implemented dose calculation after changing it to AAA, and compared and analyzed treatment plan using Dose Volume Histogram (DVH). Results: Firstly, as a result of evaluating PBC algorithm of Conventional plan and the difference according to AAA, the average difference of CI value on target volume has been highly estimated by 0.295 on PBC algorithm and as a result of evaluating dose of lung, $V_{47Gy}$ and $V_{45Gy}$ has been highly evaluated by 5.83% and 4.04% each, Mean dose, $V_{20Gy}$, $V_{5Gy}$, $V_{3Gy}$ has been highly evaluated 0.6%, 0.29%, 6.35%, 10.23% each on AAA. Secondly, in case of FiF plan, the average difference of CI value on target volume has been highly evaluated on PBC algorithm by 0.165, and dose on ipsilateral lung, $V_{47Gy}$, $V_{45Gy}$, Mean dose has been highly evaluated 6.17%, 3.80%, 0.15% each on PBC algorithm, $V_{20Gy}$, $V_{5Gy}$, $V_{3Gy}$ has been highly evaluated 0.14%, 4.07%, 4.35% each on AAA. Conclusion: When calculating with AAA on breast cancer tangential plan, compared to PBC algorithm, Conformity on target volume of Conventional plan, FiF plan has been less evaluated by 0.295, 0.165 each. For the reason that dose of high dose region of ipsilateral lung has been showed little amount, and dose of low dose region has been showed much amount, features according to dose calculation algorithm need to be considered when we evaluate dose for the lungs.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.40 no.5
• /
• pp.312-321
• /
• 2003

In this paper, we implement robot which are ability to recognize obstacles and moving automatically to destination. we present two results in this paper; hardware implementation of image processing board and software implementation of visual feedback algorithm for a self-controlled robot. In the first part, the mobile robot depends on commands from a control board which is doing image processing part. We have studied the self controlled mobile robot system equipped with a CCD camera for a long time. This robot system consists of a image processing board implemented with DSPs, a stepping motor, a CCD camera. We will propose an algorithm in which commands are delivered for the robot to move in the planned path. The distance that the robot is supposed to move is calculated on the basis of the absolute coordinate and the coordinate of the target spot. And the image signal acquired by the CCD camera mounted on the robot is captured at every sampling time in order for the robot to automatically avoid the obstacle and finally to reach the destination. The image processing board consists of DSP (TMS320VC33), ADV611, SAA7111, ADV7l76A, CPLD(EPM7256ATC144), and SRAM memories. In the second part, the visual feedback control has two types of vision algorithms: obstacle avoidance and path planning. The first algorithm is cell, part of the image divided by blob analysis. We will do image preprocessing to improve the input image. This image preprocessing consists of filtering, edge detection, NOR converting, and threshold-ing. This major image processing includes labeling, segmentation, and pixel density calculation. In the second algorithm, after an image frame went through preprocessing (edge detection, converting, thresholding), the histogram is measured vertically (the y-axis direction). Then, the binary histogram of the image shows waveforms with only black and white variations. Here we use the fact that since obstacles appear as sectional diagrams as if they were walls, there is no variation in the histogram. The intensities of the line histogram are measured as vertically at intervals of 20 pixels. So, we can find uniform and nonuniform regions of the waveforms and define the period of uniform waveforms as an obstacle region. We can see that the algorithm is very useful for the robot to move avoiding obstacles.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.35S no.12
• /
• pp.109-120
• /
• 1998

An image segmentation approach based on the fuzzy worm searching and MIN-MAX clustering algorithm is proposed in this paper. This algorithm deals with fuzzy worm value and min-max node at a gross scene level, which investigates the edge information including fuzzy worm action and spatial relationship of the pixels as the parameters of its objective function. But the conventional segmentation methods for edge extraction generally need the mask information for the algebraic model, and take long run times at mask operation, whereas the proposed algorithm has single operation according to active searching of fuzzy worms. In addition, we also propose both genetic fuzzy worm searching and genetic min-max clustering using genetic algorithm to complete clustering and fuzzy searching on grey-histogram of image for the optimum solution, which can automatically determine the size of ranges and has both strong robust and speedy calculation. The simulation results showed that the proposed algorithm adaptively divided the quantized images in histogram region and performed single searching methods, significantly alleviating the increase of the computational load and the memory requirements.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.36S no.1
• /
• pp.51-59
• /
• 1999

In this paper we propose the technique for tracking players and a ball and for obtaining players' ground traces using image mosaic in general soccer sequences. Here, general soccer sequences mean the case that there is no extreme zoom-in or zoom-out of TV camera. Obtaining player's ground traces requires that the following three main problems be solved. There main problems: (1) ground field extraction (2) player and ball tracking and team indentification (3) player positioning. The region of ground field is extracted on the basis of color information. Players are tracked by template matching and Kalman filtering. Occlusion reasoning between overlapped players in done by color histogram back-projection. To find the location of a player, a ground model is constructed and transformation between the input images and the field model is computed using four or more feature points. But, when feature points extracted are insufficient, image-based mosaic technique is applied. By this image-to-model transformation, the traces of players on the ground model can be determined. We tested our method on real TV soccer sequence and the experimental results are given.

• Convergence Security Journal
• /
• v.19 no.2
• /
• pp.91-104
• /
• 2019

In this paper, we present a novel local descriptor, Local Prominent Directional Pattern (LPDP), to represent the description of facial images for gender recognition purpose. To achieve a clearly discriminative representation of local shape, presented method encodes a target pixel with the prominent directional variations in local structure from an analysis of statistics encompassed in the histogram of such directional variations. Use of the statistical information comes from the observation that a local neighboring region, having an edge going through it, demonstrate similar gradient directions, and hence, the prominent accumulations, accumulated from such gradient directions provide a solid base to represent the shape of that local structure. Unlike the sole use of gradient direction of a target pixel in existing methods, our coding scheme selects prominent edge directions accumulated from more samples (e.g., surrounding neighboring pixels), which, in turn, minimizes the effect of noise by suppressing the noisy accumulations of single or fewer samples. In this way, the presented encoding strategy provides the more discriminative shape of local structures while ensuring robustness to subtle changes such as local noise. We conduct extensive experiments on gender recognition datasets containing a wide range of challenges such as illumination, expression, age, and pose variations as well as sketch images, and observe the better performance of LPDP descriptor against existing local descriptors.

• The Journal of Engineering Geology
• /
• v.33 no.4
• /
• pp.673-687
• /
• 2023

Exploratory data analysis is the process of observing and understanding data collected from various sources to identify their distributions and correlations through their structures and characterization. This process can be used to identify correlations among conditioning factors and select the most effective factors for analysis. This can help the assessment of landslide susceptibility, because landslides are usually triggered by multiple factors, and the impacts of these factors vary by region. This study compared two stages of exploratory data analysis to examine the impact of the data exploration procedure on the landslide prediction model's performance with respect to factor selection. Deep-learning-based landslide susceptibility analysis used either a combinations of selected factors or all 23 factors. During the data exploration phase, we used a Pearson correlation coefficient heat map and a histogram of random forest feature importance. We then assessed the accuracy of our deep-learning-based analysis of landslide susceptibility using a confusion matrix. Finally, a landslide susceptibility map was generated using the landslide susceptibility index derived from the proposed analysis. The analysis revealed that using all 23 factors resulted in low accuracy (55.90%), but using the 13 factors selected in one step of exploration improved the accuracy to 81.25%. This was further improved to 92.80% using only the nine conditioning factors selected during both steps of the data exploration. Therefore, exploratory data analysis selected the conditioning factors most suitable for landslide susceptibility analysis and thereby improving the performance of the analysis.

• Progress in Medical Physics
• /
• v.14 no.1
• /
• pp.8-14
• /
• 2003

The spinal cord dose is the one of the limiting factor for the radiation treatment of the head ＆ neck or the thorax region. It is not an easy task to maintain the spinal cord dose below tolerance and to keep the clinically acceptable dose to the PTV in this region. To overcome this problem, the spinal cord partial block technique (PBT) with the dynamic Multi-Leaf Collimator (dMLC) has been developed. This technique is an extension of the conventional treatment planning. In the beginning the beam directions are selected as same as the conventional treatment planning to encompass the PTV, then the partial block are designed to shield the spinal cord. The plan comparisons between the conventional therapy plan and the PTB plan were performed to evaluate the validity of this technique. The mean dose and the dose volume histogram (DVH) were used as the plan comparison indices. A series of quality assurance (QA) was performed to guarantee the reliable treatment. The QA consisted of the film dosimetry for the verification of the dose distribution and the point measurements. The PBT plan generated better results than the conventional treatment plan and it was proved to be useful for the H＆N region.

• Journal of Korea Multimedia Society
• /
• v.14 no.4
• /
• pp.494-503
• /
• 2011

In this paper, we propose a recognition application of facial expression for laughter theraphy on smartphone. It detects face region by using AdaBoost face detection algorithm from the front camera image of a smartphone. After detecting the face image, it detects the lip region from the detected face image. From the next frame, it doesn't detect the face image but tracks the lip region which were detected in the previous frame by using the three step block matching algorithm. The size of the detected lip image varies according to the distance between camera and user. So, it scales the detected lip image with a fixed size. After that, it minimizes the effect of illumination variation by applying the bilateral symmetry and histogram matching illumination normalization. After that, it computes lip eigen vector by using PCA(Principal Component Analysis) and recognizes laughter expression by using a multilayer perceptron artificial network. The experiment results show that the proposed method could deal with 16.7 frame/s and the proposed illumination normalization method could reduce the variations of illumination better than the existing methods for better recognition performance.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2014.10a
• /
• pp.320-323
• /
• 2014

Today is going with the advancement of intelligent transportation systems and traffic environment and helping to provide safe and convenient service through a mobile device work with the popularization of the vehicle black box. The traffic flow by a variety of causes is constantly changing, it is often unable to prepare the driver, depending on external factors can not be controlled by the power of the public, leading to a major accident. The system needs to pass the real-time weather data in the inter-operator to prevent this. The proposed detection algorithm weather information delivered real-time weather information for this paper. The weather condition is detected by using the contrast between the histogram of the motion of the wiper and the clear day algorithm. In general, the wiper is worked in extreme weather conditions that will have a value different contrast due to rain or snow. Situation was considered clear, snowy conditions, such as using it on a rainy situation. First, designated as ROI (Region Of Interest) of the minimum area that can be detected in order to reduce the amount of calculation for the wiper, the wiper, which was detected through the operation of the threshold Thresholding the brightness of the vehicle wiper. In addition, we distinguish the value of each meteorological situation by using contrast. Results was obtained to 80% for the snow conditions, a rainy situation.