• 반도체디스플레이기술학회지
• /
• 제17권2호
• /
• pp.76-80
• /
• 2018

In recent years, with the development of automobile technology, smart system technology that assists safe driving has been developed. A camera is installed on the front and rear of the vehicle as well as on the left and right sides to detect and warn of collision risks and hazards. Beyond the technology of simple black-box recording via cameras, we are developing intelligent systems that combine various computer vision technologies. However, most related studies have been developed to optimize performance in laboratory-like environments that do not take environmental factors such as weather into account. In this paper, we propose a method to detect object by restoring visibility in image with degraded image due to weather factors such as fog. First, the image quality degradation such as fog is detected in a single image, and the image quality is improved by restoring using an intermediate value filter. Then, we used an adaptive feature extraction method that removes unnecessary elements such as noise from the improved image and uses it to recognize objects with only the necessary features. In the proposed method, it is shown that more feature points are extracted than the feature points of the region of interest in the improved image.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권10호
• /
• pp.3608-3626
• /
• 2021

The detection of bolts is an important task in high-speed train inspection systems, and it is frequently performed to ensure the safety of trains. The difficulty of the vision-based bolt inspection system lies in small sample defect detection, which makes the end-to-end network ineffective. In this paper, the problem is resolved in two stages, which includes the detection network and cascaded classification networks. For small bolt detection, all bolts including defective bolts and normal bolts are put together for conducting annotation training, a new loss function and a new boundingbox selection based on the smallest axis-aligned convex set are proposed. These allow YOLOv3 network to obtain the accurate position and bounding box of the various bolts. The average precision has been greatly improved on PASCAL VOC, MS COCO and actual data set. After that, the Siamese network is employed for estimating the status of the bolts. Using the convolutional Siamese network, we are able to get strong results on few-shot classification. Extensive experiments and comparisons on actual data set show that the system outperforms state-of-the-art algorithms in bolt inspection.

• 로봇학회논문지
• /
• 제18권3호
• /
• pp.346-351
• /
• 2023

With the release of numerous open driving datasets, the demand for domain adaptation in perception tasks has increased, particularly when transferring knowledge from rich datasets to novel domains. However, it is difficult to solve the change 1) in the sensor domain caused by heterogeneous LiDAR sensors and 2) in the environmental domain caused by different environmental factors. We overcome domain differences in the semi-supervised setting with 3-stage model parameter training. First, we pre-train the model with the source dataset with object scaling based on statistics of the object size. Then we fine-tine the partially frozen model weights with copy-and-paste augmentation. The 3D points in the box labels are copied from one scene and pasted to the other scenes. Finally, we use the knowledge distillation method to update the student network with a moving average from the teacher network along with a self-training method with pseudo labels. Test-Time Augmentation with varying z values is employed to predict the final results. Our method achieved 3rd place in ECCV 2022 workshop on the 3D Perception for Autonomous Driving challenge.

• Korean Journal of Radiology
• /
• 제24권7호
• /
• pp.698-714
• /
• 2023

In recent years, artificial intelligence, especially object detection-based deep learning in computer vision, has made significant advancements, driven by the development of computing power and the widespread use of graphic processor units. Object detection-based deep learning techniques have been applied in various fields, including the medical imaging domain, where remarkable achievements have been reported in disease detection. However, the application of deep learning does not always guarantee satisfactory performance, and researchers have been employing trial-and-error to identify the factors contributing to performance degradation and enhance their models. Moreover, due to the black-box problem, the intermediate processes of a deep learning network cannot be comprehended by humans; as a result, identifying problems in a deep learning model that exhibits poor performance can be challenging. This article highlights potential issues that may cause performance degradation at each deep learning step in the medical imaging domain and discusses factors that must be considered to improve the performance of deep learning models. Researchers who wish to begin deep learning research can reduce the required amount of trial-and-error by understanding the issues discussed in this study.

• Nuclear Engineering and Technology
• /
• 제56권2호
• /
• pp.498-505
• /
• 2024

In this paper, a cable-driven endoscopic manipulator (CEM) is designed for the Chinese latest compact fusion reactor. The whole CEM arm is more than 3000 mm long and includes end vision tools, an endoscopic manipulator/control system, a feeding system, a drag chain system, support systems, a neutron shield door, etc. It can cover a range of ±45° of the vacuum chamber by working in a wrap-around mode, etc., to meet the need for observation at any position and angle. By placing all drive motors in the end drive box via a cable drive, cooling, and radiation protection of the entire robot can be facilitated. To address the CEM motion control problem, a discrete trajectory tracking method is proposed. By restricting each joint of the CEM to the target curve through segmental fitting, the trajectory tracking control is completed. To avoid the joint rotation angle overrun, a joint limit rotation angle optimization method is proposed based on the equivalent rod length principle. Finally, the CEM simulation system is established. The rationality of the structure design and the effectiveness of the motion control algorithm are verified by the simulation.

• Smart Structures and Systems
• /
• 제33권2호
• /
• pp.105-118
• /
• 2024

In order to realize tiny bridge crack discovery by UAV-based machine vision, a novel method combining deep learning and tensor voting is proposed. Firstly, the grid images of crack are detected and descripted based on SE-ResNet50 to generate feature points. Then, the probability significance map of crack image is calculated by tensor voting with feature points, which can define the direction and region of crack. Further, the crack detection anchor box is formed by non-maximum suppression from the probability significance map, which can improve the robustness of tiny crack detection. Finally, a case study is carried out to demonstrate the effectiveness of the proposed method in the Xiangjiang-River bridge inspection. Compared with the original tensor voting algorithm, the proposed method has higher accuracy in the situation of only 1-2 pixels width crack and the existence of edge blur, crack discontinuity, which is suitable for UAV-based bridge crack discovery.

• 터널과지하공간
• /
• 제34권3호
• /
• pp.218-230
• /
• 2024

최근 건설현장의 안전사고 문제를 해결하기 위해 컴퓨터 비전 기술을 활용한 안전관리에 관한 연구를 많이 수행하고 있다. 최근 딥러닝 기반 객체 인식 및 영역 분할 연구에서 앵커 박스 파라미터를 사용하고 있다. 일관적인 정확도를 확보하기 위하여 학습 과정에서 앵커 박스 파라미터의 최적화가 중요하다. 앵커 박스 관련 파라미터는 일반적으로 학습자의 휴리스틱 방법으로 모양과 크기를 고정하여 학습을 수행하고 있고, 파라미터는 단일로 구성된다. 하지만 파라미터는 객체 종류와 객체 크기에 따라 민감하고 수가 증가하면 단일 파라미터로 데이터의 모든 특성을 반영하는데 한계가 발생한다. 따라서 본 논문은 분할 학습을 통해 최적화된 다중 파라미터를 적용하는 방법을 제안하여 단일 파라미터로 모든 객체의 특성을 반영하기 어려운 문제를 해결하고자 한다. 통합 데이터를 객체 크기, 객체 수, 객체의 형상에 따라 효율적으로 분할하는 기준을 정립하였으며, 최종으로 통합 학습과 분할 학습 방법의 성능 비교를 통해 제안한 학습 방법의 효과를 검증하였다.

• 한국토양동물학회지
• /
• 제5권2호
• /
• pp.101-112
• /
• 2000

시각기가 발달되지 않은 대부분의 거미류와는 달리 잘 발달된 시각기를 갖는 배회성거미인 흰수염깡충거미 시각기의 미세구조를 광학현미경과 전자현미경으로 기본구성 형태와 구조 분석을 목적으로 관찰 조사하였다. 관찰결과 흰수염깡충거미 8개의 눈, 모두는 각막, 수정체,유리체 그리고 망막으로 이루어져 있었다. 3열로 배열된 4쌍의 시각기 중 2열에 있는 후중안이 기본구성형태와 세포의 수 그리고 크기가 작고 빈약하였으나 다른 시각기는 발달되었다. 외피의 큐티클성 각막층은 렌즈와 붙어 있었다. 유리체는 원주형세포로 이루어져 있었으며, 망막은 잘 발달된 미세융모 형태의 감간체와 비 색소지지세포 그리고 색소세포로 이루어져 있었다. 전중안은 원주형의 세포로 이루어진 유리체가 존재하였으며 유리체를 둘러싸고 있는 망막에는 빛을 감지하는 미세융모 형태의 감간체가 다른 눈들과는 달리 불규칙하게 배열되어 있었다. 부안들은 수용체세포의 횡단면절편(cross section)에서 미세융모와 비색소세포, 그리고 색소세포와 어우러져 별모양(starlike)을 나타냈다. 반사체는 볼 수 없었다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제11권6호
• /
• pp.3188-3207
• /
• 2017

In this paper, we propose an image enhancement algorithm to improve license plate extraction rate in various environments (Day Street, Night Street, Underground parking lot, etc.). The proposed algorithm is composed of image enhancement algorithm and license plate extraction algorithm. The image enhancement method can improve an image quality of the degraded image, which utilizes a histogram information and overall gray level distribution of an image. The proposed algorithm employs an interpolated probability distribution value (PDV) in order to control a sudden change in image brightness. Probability distribution value can be calculated using cumulative distribution function (CDF) and probability density function (PDF) of the captured image, whose values are achieved by brightness distribution of the captured image. Also, by adjusting the image enhancement factor of each part region based on image pixel information, it provides a function that can adjust the gradation of the image in more details. This processed gray image is converted into a binary image, which fuses narrow breaks and long thin gulfs, eliminates small holes, and fills gaps in the contour by using morphology operations. Then license plate region is detected based on aspect ratio and license plate size of the bound box drawn on connected license plate areas. The images have been captured by using a video camera or a personal image recorder installed in front of the cars. The captured images have included several license plates on multilane roads. Simulation has been executed using OpenCV and MATLAB. The results show that the extraction success rate is more improved than the conventional algorithms.

• 복식
• /
• 제63권5호
• /
• pp.151-166
• /
• 2013

In order to create a national myth and be able to control international society, America with her short national history, used popular culture to accomplish these goals. The medium fit this purpose the best was the use of superhero characters based on comics. Born and developed from the 1930s through the 1960s, which could be seen as America's national crisis era, superhero characters were thorough advocates of American justice and was perfect for the role of spreading the legitimacy of American ideology. From the 1970s, superhero characters became part of movies and became even more influential through the Hollywood's massive film industry and the box office success. American ideology in superhero characters symbolically appeared in movie costumes. Starting with Superman and Batman, the very first and typical superhero characters' costumes work as metaphors for realization of American justice. After the 1980s, superheroes were newly developed through a genre called graphic novel and the most representative piece of this genre is Alan Moore's Watchmen. In the Watchmen, which was also turn into a movie in 2009, six changed superhero characters appear ranging from a non-human superhero, villain superhero, superhero with mental disorder and superhero with sexual impotency, the characters were never-seen-before superheroes with different aspects that connote introspection and philosophical ideology. The changed type of heroes and ideology became another form of heroes, and this brought changes to character costumes that were never considered before. The superhero costumes that used to symbolize America now express different types of superhero by borrowing exotic mythical elements, undressing, pastiche and daily life clothes. The superhero characters and their changes in costumes from Watchmen imply American popular culture's introspective tendency. Amongst these changes, we need to raise our critical vision towards popular culture.