• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2022년도 추계학술발표대회
• /
• pp.625-627
• /
• 2022

In the construction and development of modern industrial production technology, the traditional technology management mode is faced with many problems such as low qualification rates and high application costs. In the research, an improved workpiece defect detection method based on deep learning is proposed, which can control the application cost and improve the detection efficiency of irregular defects. Based on the research of the current situation of deep learning applications, this paper uses the improved Faster R-CNN network structure model as the core detection algorithm to automatically locate and classify the defect areas of the workpiece. Firstly, the robustness of the model was improved by appropriately changing the depth and the number of channels of the backbone network, and the hyperparameters of the improved model were adjusted. Then the deformable convolution is added to improve the detection ability of irregular defects. The final experimental results show that this method's average detection accuracy (mAP) is 4.5% higher than that of other methods. The model with anchor size and aspect ratio (65,129,257,519) and (0.2,0.5,1,1) has the highest defect recognition rate, and the detection accuracy reaches 93.88%.

• 한국컴퓨터정보학회논문지
• /
• 제21권10호
• /
• pp.11-19
• /
• 2016

The size of a display is large, The form becoming various of that do not apply to previous methods of gaze tracking and if setup gaze-track-camera above display, can solve the problem of size or height of display. However, This method can not use of infrared illumination information of reflected cornea using previous methods. In this paper, Robust pupil detecting method for eye's occlusion, corner point of inner eye and center of pupil, and using the face pose information proposes a method for calculating the simply position of the gaze. In the proposed method, capture the frame for gaze tracking that according to position of person transform camera mode of wide or narrow angle. If detect the face exist in field of view(FOV) in wide mode of camera, transform narrow mode of camera calculating position of face. The frame captured in narrow mode of camera include gaze direction information of person in long distance. The method for calculating the gaze direction consist of face pose estimation and gaze direction calculating step. Face pose estimation is estimated by mapping between feature point of detected face and 3D model. To calculate gaze direction the first, perform ellipse detect using splitting from iris edge information of pupil and if occlusion of pupil, estimate position of pupil with deformable template. Then using center of pupil and corner point of inner eye, face pose information calculate gaze position at display. In the experiment, proposed gaze tracking algorithm in this paper solve the constraints that form of a display, to calculate effectively gaze direction of person in the long distance using single camera, demonstrate in experiments by distance.

• Geomechanics and Engineering
• /
• 제14권3호
• /
• pp.283-291
• /
• 2018

Cracks in soil provide significant preferential pathways for contaminant transport and rainfall infiltration. Water exchange between the soil matrix and crack is crucial to characterize the preferential flow, which is often quantitatively described by a water exchange ratio. The water exchange ratio is defined as the amount of water flowing from the crack into the clay matrix per unit time. Most of the previous studies on the water exchange ratio mainly focused on cracked sandy soils. The water exchange between cracks and clay matrix were rarely studied mainly due to two reasons: (1) Cracks open upon drying and close upon wetting. The deformable cracks lead to a dynamic change in the water exchange ratio. (2) The aperture of desiccation crack in clay is narrow (generally 0.5 mm to 5 mm) which is difficult to model in experiments. This study will investigate the water exchange between a deformable crack and the clay matrix using a newly developed experimental apparatus. An artificial crack with small aperture was first fabricated in clay without disturbing the clay matrix. Water content sensors and suction sensors were instrumented at different places of the cracked clay to monitor the water content and suction changes. Results showed that the water exchange ratio was relatively large at the initial stage and decreased with the increasing water content in clay matrix. The water exchange ratio increased with increasing crack apertures and approached the largest value when the clay was compacted at the water content to the optimal water content. The effective hydraulic conductivity of the crack-clay matrix interface was about one order of magnitude larger than that of saturated soil matrix.

• 한국컴퓨터그래픽스학회논문지
• /
• 제15권3호
• /
• pp.1-9
• /
• 2009

최근 역감, 촉감 피드백을 생성할 수 있고 3 차원 상호작용이 가능한 컴퓨터 인터페이스 기술이 발전하면서 기존의 입출력 장치에 의존하던 컴퓨터 시뮬레이션의 한계가 줄어들어 그 종류가 다양해지고 있다. 특히, 가상현실 기반 수술 시뮬레이터의 수요가 증가 하면서 사실적인 가상환경을 제공하는 방법에 관련된 연구가 활발히 진행 되고 있으며 그 중에서 가상 장기 모델의 사실적인 표현 기술은 사용자의 몰입감을 높여주는데 중요한 역할을 하기 때문에 반드시 필요한 분야이다. 가상 장기 모델의 사실성을 높이기 위해서는 사용자의 상호작용에 따라 변형되는 시각적 피드백과 알맞은 햅틱 피드백을 전달해야한다. 본 연구에서는 이 두 가지 피드백은 사람이 인지하는 방법과 시스템의 요구 사항이 다르기 때문에 보다 효과적인 시뮬레이션을 하기 위하여 서로 다른 두 모델을 구현하고 두 모델간의 동기화를 위한 기술로써 적응성 질량-스프링 모델 기법을 제안한다. 또한, 유연체 변형에 대하여 표면의 세부 형태를 보여주는 시각모델과 사용자의 상호작용에 따른 햅틱 피드백을 전달하는 햅틱 모델, 그리고 이 두개의 모델을 연동시키는 동기화 모델을 통하여 유연체의 거동을 실시간, 사실적으로 제공할 수 있는 가변형 모델 시뮬레이션 프레임워크를 설계한다.

• 대한의용생체공학회:의공학회지
• /
• 제28권6호
• /
• pp.713-720
• /
• 2007

In image-guided surgery, automatic bone segmentation of Computed Tomography (CT) images is an important but challenging step. Previous attempts include intensity-, edge-, region-, and deformable curve-based approaches [1], but none claims fully satisfactory performance. Although active contour (AC) techniques possess many excellent characteristics, their applications in CT image segmentation have not worthily exploited yet. In this study, we have evaluated the automaticity and performance of the model of Chan-Vese Multiphase AC Without Edges towards knee bone segmentation from CT images. This model is suitable because it is initialization-insensitive and topology-adaptive. Its segmentation results have been qualitatively compared with those from four other widely used AC models: namely Gradient Vector Flow (GVF) AC, Geometric AC, Geodesic AC, and GVF Fast Geometric AC. To quantitatively evaluate its performance, the results from a commercial software and a medical expert have been used. The evaluation results show that the Chan-Vese model provides superior performance with least user interaction, proving its suitability for automatic bone segmentation from CT images.

• 대한기계학회논문집A
• /
• 제25권4호
• /
• pp.582-591
• /
• 2001

It is known that the reduced integration, modified shape function, anisoparametric and non-conforming element can reduce the error induced by stiffness locking phenomenon in the finite element analysis. In this study, we propose new anisoparametric curved beam element. The new element based on reduced minimization theory is composed of different shape functions in each displacement field. By the substitution of this modified shape function, the unmatched coefficient that cause stiffness locking in the constraint energy is eliminated. To confirm the availability of this new model, we performed numerical tests for a simple model. As a result of numerical test, the undulate stress patterns are disappeared in static analysis, and displacements and stresses are close to exact solution. Not only in the static analysis but also in the eigen analysis of free vibrated curved beam model, this element shows successful convergent results.

• 한국ITS학회 논문지
• /
• 제2권2호
• /
• pp.85-95
• /
• 2003

본 논문은 Snake 모델을 이용하여 동영상에서 주위 환경 변화에 적응 가능한 다중 이동 객체 추적 시스템을 제안하였다. Snake 모델은 배경이 복잡한 영상에 대해선 객체의 윤곽선을 정확히 표현하지 못하므로 영상분할 시 초기 위치에 따라 민감하게 영향을 받는다. 제안된 시스템은 프레임간의 차(difference)영상을 이용하여 배경영상을 획득하고, 픽셀의 인접성을 조사하여 객체를 분할하고 위치 특징 값을 구하며, 분할된 특징 값들을 Snake모델의 초기 위치 값으로 부여함으로써 초기 위치 값에 민감한 Snake 모델을 개선하였다 또한 본 시스템은 복잡한 배경 영상을 단순화하고, Snake를 이루는 각 정점들을 객체의 위치로 놓이게 함으로써 탐색 공간을 줄였다. 30fps로 저장된 AVI파일을 적용함으로써 다중 이동차량 추적 시스템으로의 응용 가능함을 보였다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2003년도 하계종합학술대회 논문집 Ⅳ
• /
• pp.1968-1971
• /
• 2003

We propose a knowledge-based algorithm for extracting an object boundary from low-quality image like the forward looking infrared image. With the multi-classes training data set, the global shape is modeled by multispace KL(MKL)[1] and curvature model. And the objective function for fitting the deformable boundary template represented by the shape model to true boundary in an input image is formulated by Bales rule. Simulation results show that our method has more accurateness in case of multi-classes training set and performs better in the sense of computation cost than point distribution model(PDM)[2]. It works well in distortion under the noise, pose variation and some kinds of occlusions.

• Steel and Composite Structures
• /
• 제25권5호
• /
• pp.625-638
• /
• 2017

This paper presents a novel approach that describes the first-order (linear elastic) partial interaction analysis of members formed by multi-components based on the Generalised Beam Theory (GBT). The novelty relies on its ability to accurately model the partial interaction between the different components forming the cross-section in both longitudinal and transverse directions as well as to consider the cross-sectional deformability. The GBT deformations modes, that consist of the conventional, extensional and shear modes, are determined from the dynamic analyses of the cross-section represented by a planar frame. The partial interaction is specified at each connection interface between two adjacent elements by means of a shear deformable spring distributed along the length of the member. The ease of use of the model is outlined by an application performed on a multi-component member subjected to an eccentric load. The values calculated with an ABAQUS finite element model are used to validate the proposed method. The results of the numerical applications outline the influence of specifying different rigidities for the interface shear connection and in using different order of polynomials for the shape functions specified in the finite element cross-section analysis.

• 자동차안전학회지
• /
• 제7권1호
• /
• pp.27-32
• /
• 2015

In order to protect occupant during car crash accident, Regulation and NCAP(New Car assessment Program) have been developed among various countries like U.S.A., Europe, Korea and Australia. Especially NCAP scores affect to sales of vehicles. So vehicle makers are trying to get good score in NCAP. Low leg injuries play an important role in Australia and Euro NCAP and these injuries are related with footrest design. Optimization of footrest design in early stage of vehicle development is necessary to obtain better and robust results of low legs during crash tests. In this paper, DFSS method and finite element model were used to optimize the low leg performance in small RHD vehicles. Compared with the lower leg injury of base model, the lower leg injury of proposed model was slightly improved and robustness was enhanced also.