• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.20 no.4
• /
• pp.95-105
• /
• 2021

Road surface damage detection is essential for a comfortable driving environment and the prevention of safety accidents. Road management institutes are using automated technology-based inspection equipment and systems. As one of these automation technologies, a sensor to detect road surface damage plays an important role. For this purpose, several studies on sensors using deep learning have been conducted in recent years. Road images and label images are needed to develop such deep learning algorithms. On the other hand, considerable time and labor will be needed to secure label images. In this paper, the adversarial learning method, one of the semi-supervised learning techniques, was proposed to solve this problem. For its implementation, a lightweight deep neural network model was trained using 5,327 road images and 1,327 label images. After experimenting with 400 road images, a model with a mean intersection over a union of 80.54% and an F1 score of 77.85% was developed. Through this, a technology that can improve recognition performance by adding only road images was developed to learning without label images and is expected to be used as a technology for road surface management in the future.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.2402-2406
• /
• 2005

A collision query determines the intersection between given objects, and is used in computer-aided design and manufacturing, animation and simulation systems, and physically-based modeling. Bounding volume hierarchies are one of the simplest and most widely used data structures for performing collision detection on complex models. In this paper, we present hierarchy of oriented rounded bounding volume for fast proximity queries. Designing hierarchies of new bounding volumes, we use to combine multiple bounding volume types in a single hierarchy. The new bounding volume corresponds to geometric shape composed of a core primitive shape grown outward by some offset such as the Minkowski sum of rectangular box and a sphere shape. In the experiment of parallel close proximity, a number of benchmarks to measure the performance of the new bounding box and compare to that of other bounding volumes.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.11 no.7
• /
• pp.641-648
• /
• 2001

A face detection in color images is important for many multimedia applications. It is first step for face recognition and can be used for classifying specific shorts. This paper describes a new method to detect faces in color images based on the skin color and hair color. This paper presents a fuzzy-based method for classifying skin color region in a complex background under varying illumination. The Fuzzy rule bases of the fuzzy system are generated using training method like a genetic algorithm(GA). We find the skin color region and hair color region using the fuzzy system and apply the convex-hull to each region and find the face from their intersection relationship. To validity the effectiveness of the proposed method, we make experiment with various cases.

• Korean Journal of Computational Design and Engineering
• /
• v.4 no.4
• /
• pp.312-317
• /
• 1999

An inflection point on a curve is a point where the curvature vanishes. An inflection point is useful for various geometric operations such as the approximation of curves and intersection points between curves or curve approximations. An inflection point on planar Bezier curves can be easily detected using a hodograph and a derivative of hodograph, since the closed from of hodograph is known. In the case of rational Bezier curves, for the detection of inflection point, it is needed to use the first and the second derivatives have higher degree and are more complex than those of non-rational Bezier curvet. This paper presents three methods to detect inflection points of rational Bezier curves. Since the algorithms avoid explicit derivations of the first and the second derivatives of rational Bezier curve to generate polynomial of relatively lower degree, they turn out to be rather efficient. Presented also in this paper is the theoretical analysis of the performances of the algorithms as well as the experimental result.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.3
• /
• pp.471-477
• /
• 1996

An image edge detection method based on the toll set concept is proposed. Initially the edge structure is established for an image following human perception n model. Then toll set membership values are computed and the toll set intersection and union operators are applied to them. The final toll set membership values are normalized to get the vagueness degrees and the thresholding operation based on entropy concept is performed on them to determine the edge of an image.

• Smart Structures and Systems
• /
• v.32 no.2
• /
• pp.111-121
• /
• 2023

This study proposes an automated assembly performance evaluation method for prefabricated steel structures (PSSs) using machine learning methods. Assembly component images were segmented using a modified version of the receptive field pyramid. By factorizing channel modulation and the receptive field exploration layers of the convolution pyramid, highly accurate segmentation results were obtained. After completing segmentation, the positions of the bolt holes were calculated using various image processing techniques, such as fuzzy-based edge detection, Hough's line detection, and image perspective transformation. By calculating the distance ratio between bolt holes, the assembly performance of the PSS was estimated using the k-nearest neighbors (kNN) algorithm. The effectiveness of the proposed framework was validated using a 3D PSS printing model and a field test. The results indicated that this approach could recognize assembly components with an intersection over union (IoU) of 95% and evaluate assembly performance with an error of less than 5%.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2024.01a
• /
• pp.73-75
• /
• 2024

본 논문에서는 밀 수확량을 증가시키기 위한 일반화된 검출 모델을 제안한다. 일반화 성능을 높이기 위해 CutMix 알고리즘으로 데이터를 증식시켰고, 라벨링 되지 않은 데이터를 최대한 활용하기 위해 Fast R-CNN 기반 Pseudo labeling을 사용하였다. 학습의 정확성과 효율성을 높이기 위해 사전에 훈련된 EfficientDet 모델로 학습하였으며, OOF를 이용하여 검증하였다. 최신 객체 검출 모델과 IoU(Intersection over Union)를 이용한 성능 평가 결과, 제안된 모델이 가장 높은 성능을 보이는 것을 확인하였다.

• The Journal of the Convergence on Culture Technology
• /
• v.4 no.4
• /
• pp.367-372
• /
• 2018

This paper proposed an algorithm that detects boundary nodes effectively in wireless sensor network. A boundary node is a sensor that lies on the border of network holes or the outer boundary of wireless sensor network. Proposed algorithm detects boundary nodes using only the position information of sensors. In addition, to improve detect performance, sensor computes the overlap area of nearest sensor first. Simulation is performed to validate the process of the proposed algorithm. In Simulation, several obstacles are placed and varying number of sensors in the range of 500~1500 are deployed in the area in order to reflect real world. The simulation results shows that proposed algorithm detects boundary nodes effectively that are located on the border of holes and the outer boundary of wireless sensor network.

• KIPS Transactions on Software and Data Engineering
• /
• v.7 no.1
• /
• pp.25-32
• /
• 2018

Due to heavy computational cost, deformable object simulation requires more effective collision detection method than rigid body simulation. However, when the CPU-based collision detection algorithm is purely applied to the GPU environment, the collision detection algorithm and the data structure optimized for the GPU environment are essential because the performance of the GPU can not be used properly. Therefore, we propose a GPU-based parallel collision detection algorithm for mass-spring system which is widely used for deformable object representation in this paper. The proposed method uses a parallel algorithm and data structure to reduce collision detection cost through GPU-based curling algorithm using AABB-Octree structure. In this paper, we prove the effectiveness of the proposed method by comparing the intersection test of all triangle pairs in parallel. The results of experimental tests show that the proposed method improves the performance by about 24% on average. Therefore, it is expected that the proposed method can improve the performance of real-time simulation for deformable objects.

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.1090-1093
• /
• 2004

In this paper, a problem of geometrically distorted images is considered. In particular, the paper discusses the detection of a watermark from a photographed image of the watermarked picture. The image is possibly obtained by using a digital camera. This watermark detection problem is made difficult by various geometric distortions added to the original picture through the printing and photographing processes. In particular, the paper focuses on the geometric distortion due to a projective transformation, as part of a camera 3D-to-2D imaging process. It is well-known that a cross ratio of collinear points is invariant under a perspective projection. By exploiting this fact, a projective-invariant digital watermarking technique is developed. By detecting the picture's corners, and the image center point at the intersection of two main diagonal lines, predefined cross ratios are used to compute the watermark embedded locations. From those identified embedding pixel locations, a watermark can be detected by performing a correlation between a watermark pattern and the image over those pixels. The proposed method does not require an inverse transformation on the distorted image, thus simplifying the detection process. Performance of the proposed method has been analyzed through computer experiments