• Journal of the Korea Computer Graphics Society
• /
• v.26 no.3
• /
• pp.31-37
• /
• 2020

Identification of the mandibular canal path in Computed Tomography (CT) scans is important in dental implantology. Typically, prior to the implant planning, dentists find a sagittal plane where the mandibular canal path is maximally observed, to manually identify the mandibular canal. However, this is time-consuming and requires extensive experience. In this paper, we propose a deep-learning-based framework to detect the desired sagittal plane automatically. This is accomplished by utilizing two main techniques: 1) a modified version of the iterative transformation network (ITN) method for obtaining initial planes, and 2) a fine searching method based on a convolutional neural network (CNN) classifier for detecting the desirable sagittal plane. This combination of techniques facilitates accurate plane detection, which is a limitation of the stand-alone ITN method. We have tested on a number of CT datasets to demonstrate that the proposed method can achieve more satisfactory results compared to the ITN method. This allows dentists to identify the mandibular canal path efficiently, providing a foundation for future research into more efficient, automatic mandibular canal detection methods.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.18 no.3
• /
• pp.697-705
• /
• 2014

This paper presents a vision based object width measurement method and its application where the light sectioning method is employed. The target object for measurement is a tread, which is the most outside component of an automobile tire. The entire system applying the measurement method consists of two processes, i.e. a calibration process and a detection process. The calibration process is to identify the relationships between a camera plane and a laser plane, and to estimate a camera lens distortion parameters. As the process requires a test pattern, namely a jig, which is elaborately manufactured. In the detection process, first of all, the region that a laser light illuminates is extracted by applying an adaptive thresholding technique where the distribution of the pixel brightness is considered to decide the optimal threshold. Then, a thinning algorithm is applied to the region so that the ends and the shoulders of a tread are detected. Finally, the tread width and the shoulder width are computed using the homography and the distortion coefficients obtained by the calibration process.

• Journal of the Korea Society of Computer and Information
• /
• v.6 no.2
• /
• pp.14-19
• /
• 2001

This paper presents a new method of 3-D surface feature extraction using a quadratic pol expression. With a range image, we get an edge map through the modified scan line technique this edge map, we label a 3-dimensional object to divide object's region and extract cent corner points from it's region. Then we determine whether the segmented region is a planar or a curved from the quadric surface equation. we calculate the coefficients of the planar su the curved surface to represent regions. In this article. we prove performance of the metho synthetic and real (Odetics) range images.

• Explosives and Blasting
• /
• v.35 no.1
• /
• pp.18-26
• /
• 2017

In Analysis of Fragmentation of rock blasted, The photo analysis method has been mainly used and these image acquisitions are mainly obtained by digital image from the front of the crushed rock. However, Image analysis is basically advantage of the image of planar shooting not front shooting but There is no way to take a photograph of huge plane rock slope. Thus, Unavoidably It is resolved by distorting or extending the image filmed at the front as well as adjusting it similar to its angle of plane shooting. Lately, With advancing unmanned aerial vehicle, It can simply image the fragment conditions of blasted rock of a high-definition digital image and Through it, It can acquire the most planar image to angle which accumulate cataclastic rock and also can make image analysis. In this study, It has been confirmed that tolerance value of analysis result of image filmed flatly is markedly lower than the existing front filmed image.

• Journal of the Korean Society of Radiology
• /
• v.17 no.4
• /
• pp.523-529
• /
• 2023

In order to achieve excellent spatial resolution, very small scintillation pixels are used in detectors of positron emission tomography for small animals. However, by using these very small scintillation pixels, scintillation pixels at the edge of the array may overlap in a flood image. To solve this problem, a light guide capable of changing the distribution of light was used. Depending on the material of the light guide, the light spreading tendency is different, and accordingly, the presence or absence of overlapping is different depending on the material of the light guide used. In this study, instead of the conventional glass light guide, a detector using the same material as the scintillation pixel was designed. A scintillator light guide has a higher refractive index than a glass light guide, so the light spread is different. Flood images were acquired to evaluate the degree of separation of the scintillation pixels at the edge of the detector using the two light guides. The degree of separation was evaluated by calculating the distance between the center and the spatial resolution of the image of two scintillation pixels at the edge of the obtained flood image. As a result, when the scintillator light guide was used, better spatial resolution was shown, and the distance between centers of scintillation pixels was wider. When a detector is constructed using a scintillator light guide instead of a conventional glass light guide, it is possible to use a smaller scintillation pixel, thereby securing better spatial resolution.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2005.11a
• /
• pp.229-232
• /
• 2005

컴퓨터 비전을 기반으로 하는 인간과 컴퓨터의 상호작용(Human computer Interaction, HCI)에 대한 연구는 영상처리 분야에서 큰 축을 담당하고 있으며, 특히 얼굴인식 연구는 HCI 분야에서 가장 중요한 영역들 중의 분야이다. 이러한 얼굴인식 기반의 HCI 시스템을 구현하기 위해서는 영상 내에 존재하는 얼굴을 정확히 검증하는 것이 선행되어야 한다. 본 논문에서는 피부색상과 원형 투영 과정에 의한 특징 추출을 이용한 특징점 기반의 얼굴 검출 기법을 제안한다. 본 논문에서 제안하는 얼굴검출 기법은 얼굴의 크기 및 평면적 회전(rotation)에 대하여 강인한 얼굴검출 성능을 보여준다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.2
• /
• pp.279-284
• /
• 2010

The plane curve approach which was proposed by Boukharouba et. al. is a multi-threshold selection method through searching peak-valley based on histogram cumulative distribution function. However the method is required to select parameters to compose plane curve, and the shape of plane curve is affected according to parameters. Therefore detection of peak-valley is effected by parameters. In this paper, we propose an entropy maximizing-based method to select optimal plane curve parameters, and propose a multi-thresholding method based on the selected parameters. The effectiveness of the proposed method is demonstrated by multi-thresholding experiments on various images and comparison with other conventional thresholding methods based on histogram.

• Journal of the Korean Society of Radiology
• /
• v.16 no.2
• /
• pp.157-162
• /
• 2022

Since preclinical positron emission tomography imaging is performed on small animals that are very small compared to the human body, a detector with excellent spatial resolution is required. For this purpose, a system was constructed using a detector using small scintillation pixels. Since the size of the currently developed and used photosensors is limited, excellent spatial resolution can be obtained when the minimum scintillation pixel and maximum array are used. In this study, the size of the photosensor is fixed and various scintillation pixel arrays are configured to match the size of the scintillation pixels, so that no overlap occurs in the flood image and the maximum scintillation pixel array in which all scintillation pixels are distinguished. For this purpose, DETECT2000, which can simulate a detector module composed of a scintillator and an photosensor, was used. A photosensor consisting of a 4 × 4 array of 3 mm × 3 mm pixels was used, and the scintillation pixel array was configured from 8 × 8 to 13 × 13, and simulations were performed. A flood image was constructed using the data obtained from the photosensor pixel, and the maximum scintillation pixel array that does not overlap the image was found through the flood image and the profile. As a result, the size of the scintillation pixel array in which all scintillation pixels are imaged without overlapping each other in the flood image was 11 × 11.

• 한국HCI학회:학술대회논문집
• /
• 2008.02a
• /
• pp.795-800
• /
• 2008

카메라로 입력되는 영상에서 객체를 인식하기 위한 노력은, 기존의 컴퓨터 비전분야에서 좋은 이슈로 연구되고 있다. 영상 내부에 등장하는 객체를 인식하고 해당 객체를 포함하고 있는 전체 이미지에서 현재 영상의 위치를 인식하기 위해서는, 영상 내에 등장할 객체에 대한 트레이닝이 필요하다. 본 논문에서는 영상에 등장할 객체에 대해서, 특징 점을 검출(feature detection)하고 각 점들이 가지는 픽셀 그라디언트 방향의 벡터 값들을 그 이웃하는 벡터 값들과 함께 DoG(difference-of-Gaussian)함수를 이용하여 정형화 한다. 이는 추후에 입력되는 영상에서 검출되는 특징 점들과 그 이웃들 간의 거리나 스케일의 비율 등의 파리미터를 이용하여 비교함으로써, 현재 특징 점들의 위치를 추정하는 정보로 사용된다. 본 논문에서는 광역의 시설 단지를 촬영한 인공위성 영상을 활용하여 시설물 내부에 존재는 건물들에 대한 초기 특징 점들을 검출하고 데이터베이스로 저장한다. 트레이닝이 마친 후에는, 프린트된 인공위성 영상내부의 특정 건물을 카메라를 이용하여 촬영하고, 이 때 입력된 영상의 특징 점을 해석하여 기존에 구축된 데이터베이스 내의 특징 점과 비교하는 과정을 거친다. 매칭되는 특징 점들은 DoG로 정형화된 벡터 값들을 이용하여 해당 건물에 대한 위치를 추정하고, 3차원으로 기 모델링 된 건물을 증강현실 기법을 이용하여 영상에 정합한 후 가시화 한다.

• Journal of the Korean Society of Radiology
• /
• v.18 no.2
• /
• pp.65-71
• /
• 2024

To achieve high resolution and sensitivity of positron emission tomography (PET) for small animals, the detector is constructed using very thin and long scintillation pixels. Due to the structure of these scintillation pixels, spatial resolution deterioration occurs outside the system's field of view. To solve this problem, we designed a detector that could improve spatial resolution by measuring the interaction depth and improve sensitivity by using a quasi-block scintillator. A quasi-block scintillator size of 12.6 mm x 12.6 mm x 3 mm was arranged in four layers, and optical sensors were placed on all sides to collect light generated by the interaction between gamma rays and the scintillator. DETECT2000 simulation was performed to evaluate the performance of the designed detector. Flood images were acquired by generating gamma-ray events at 1 mm intervals from 1.3 mm to 11.3 mm within the scintillator of each layer. The spatial resolution and peak-to-peak distance for each location were measured in an 11 x 11 array of flood images. The average measured spatial resolution was 0.25 mm, and the average distance between peaks was 1.0 mm. Through this, it was confirmed that all locations were separated from each other. In addition, because the light signals of all layers were measured separately from each other, the layer of the scintillator that interacted with the gamma rays could be completely separated. When the designed detector is used as a detector in a PET system for small animals, it is considered that excellent spatial resolution and sensitivity can be achieved and image quality can be improved.