• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.07a
• /
• pp.539-540
• /
• 2023

최근 물체 인식 모델의 성능을 개선하기 위한 다양한 연구가 진행 중이다. 본 논문에서는 K-means 기반 앵커박스 선정 기법을 적용한 새로운 물체 인식 모델 성능 개선 방법을 제안한다. 제안된 방법은 항만 내 설치된 컨테이너 사고를 예방하기 위한 컨테이너 사고위험도 분류 모델에 적용하여 성능 평가를 하였다. 특히, 컨테이너 사고위험도 분류 모델은 작은 물체를 인식해야 하며 이런 환경에서는 기존 물체 인식 모델 성능이 낮게 나타난다. 본 논문에서는 제안한 K-means 기반 앵커박스 선정 기법을 적용하여 물체 인식 모델 성능이 개선됨을 확인하였디.

• Journal of radiological science and technology
• /
• v.33 no.1
• /
• pp.31-36
• /
• 2010

Low energy high resolution (LEHR) collimator is the most widely used collimator in SPECT imaging. LEHR has an advantage in terms of image resolution but has a difficulty in acquiring high sensitivity due to the narrow hole size and long septa height. Throughput in SPECT can be improved by increasing counts per second with the use of high sensitivity collimators. The purpose of this study is to develop a system model in iterative image reconstruction to recover the resolution degradation caused by high sensitivity collimators with bigger hole size. We used fan-beam model instead of parallel-beam model for calculation of detection probabilities to accurately model the high sensitivity collimator with wider holes. In addition the weight factors were calculated and applied onto the probabilities as a function of incident angle of incoming photons and distance from source to the collimator surface. The proposed system model resulted in the equivalent performance with the same counts (i.e. in shortened acquisition time) and improved image quality in the same acquisition time. The proposed method can be effectively applied for resolution improvement of pixel collimator of next generation solid state detectors.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.16 no.5
• /
• pp.331-337
• /
• 2023

In this paper, we propose a deep learning model to detect lumbar 3 (L3) CT images to determine the occurrence and degree of sarcopenia. In addition, we would like to propose an optimization technique that uses oversampling ratio and class weight as design parameters to address the problem of performance degradation due to data imbalance between L3 level and non-L3 level portions of CT data. In order to train and test the model, a total of 150 whole-body CT images of 104 prostate cancer patients and 46 bladder cancer patients who visited Gangneung Asan Medical Center were used. The deep learning model used ResNet50, and the design parameters of the optimization technique were selected as six types of model hyperparameters, data augmentation ratio, and class weight. It was confirmed that the proposed optimization-based L3 level extraction model reduced the median L3 error by about 1.0 slices compared to the control model (a model that optimized only 5 types of hyperparameters). Through the results of this study, accurate L3 slice detection was possible, and additionally, we were able to present the possibility of effectively solving the data imbalance problem through oversampling through data augmentation and class weight adjustment.

• Bulletin of the Korean Space Science Society
• /
• 2002.10a
• /
• pp.43.2-43
• /
• 2002

• Bulletin of the Korean Space Science Society
• /
• 2003.05b
• /
• pp.34-34
• /
• 2003

• Proceedings of the Korean Nuclear Society Conference
• /
• 1998.10a
• /
• pp.83-83
• /
• 1998

• Proceedings of the Korean Institute of Communication Sciences Conference
• /
• 1996.06a
• /
• pp.102-106
• /
• 1996

• Proceedings of the Korean Radioactive Waste Society Conference
• /
• 2016.10a
• /
• pp.467-468
• /
• 2016

• Journal of Sensor Science and Technology
• /
• v.11 no.5
• /
• pp.263-272
• /
• 2002

In this paper, we investigate the potential for retrieval of morphometric data from three dimensional images of conducting bronchus obtained by X-ray Computerized Tomography (CT) and to explore the potential for the use of rapid prototype machine to produce physical hollow bronchus casts for mathematical modeling and experimental verification of particle deposition models. We segment the bronchus of lung by mathematical morphology method from obtained images by CT. The surface data representing volumetric bronchus data in three dimensions are converted to STL(streolithography) file and three dimensional solid model is created by using input STL file and rapid prototype machine. Two physical hollow cast models are created from the CT images of bronchial tree phantom and living human bronchus. We evaluate the usefulness of the rapid prototype model of bronchial tree by comparing diameters of the cross sectional area bronchus segments of the original CT images and the rapid prototyping-derived models imaged by X-ray CT.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.39 no.6
• /
• pp.49-58
• /
• 2002

Based on electrophysiological retina mechanism, a retina model is proposed, which has similar response characteristics compared with the real primate retina. Photoreceptors, horizontal cells, and bipolar cells are modeled based on the previously studied retina models. And amacrine cells known to have relation to movements detection, and bipolar cell terminals are newly modeled using 3 NDP mechanism. The proposed model verified by analyzing the spatial response characteristics to stationary and moving stimuli, and characteristics for different speeds. Through this retina model, human vision system could be applied to computer vision systems for movement detection, and it could be the basic research for the implantable artificial retina.