• Proceedings of the KIEE Conference
• /
• 1999.07g
• /
• pp.2995-2997
• /
• 1999

Recently, real-time visual tracking control for a robot manipulator is performed by using a vision feedback sensor information. In this paper, the optical flow is computed based on the eye-in-hand robot configuration. The image jacobian is employed to calculate the rotation and translation velocity of a 3D moving object. LQG visual controller generates the real-time visual trajectory. In order to improving the visual tracking performance. VSC controller is employed to control the robot manipulator. Simulation results show a better visual tracking performance than other method.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.16 no.5
• /
• pp.179-186
• /
• 2021

For stereo matching based on deep learning, the design of network structure is crucial to the calculation of matching cost, and the time-consuming problem of convolutional neural network in image processing also needs to be solved urgently. In this paper, a method of stereo matching using sparse loss volume in parallax dimension is proposed. A sparse 3D loss volume is constructed by using a wide step length translation of the right view feature map, which reduces the video memory and computing resources required by the 3D convolution module by several times. In order to improve the accuracy of the algorithm, the nonlinear up-sampling of the matching loss in the parallax dimension is carried out by using the method of multi-category output, and the training model is combined with two kinds of loss functions. Compared with the benchmark algorithm, the proposed algorithm not only improves the accuracy but also shortens the running time by about 30%.

• The Bulletin of The Korean Astronomical Society
• /
• v.44 no.1
• /
• pp.82.3-82.3
• /
• 2019

In this study, we generate future full disk magnetograms in 12, 24, 36 and 48 hours advance from SDO/HMI images using deep learning. To perform this generation, we apply the convolutional generative adversarial network (cGAN) algorithm to a series of SDO/HMI magnetograms. We use SDO/HMI data from 2011 to 2016 for training four models. The models make AI-generated images for 2017 HMI data and compare them with the actual HMI magnetograms for evaluation. The AI-generated images by each model are very similar to the actual images. The average correlation coefficient between the two images for about 600 data sets are about 0.85 for four models. We are examining hundreds of active regions for more detail comparison. In the future we will use pix2pix HD and video2video translation networks for image prediction.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2018.06a
• /
• pp.347-348
• /
• 2018

본 논문에서는 야간 도로 영상을 보정하여 주간 영상으로 변환하는 알고리즘을 제안한다. 영상 변환 딥러닝 알고리즘인 Generative Adversarial Network(GAN)를 기반으로 주야간 도로 영상을 학습시켜 주야간 상호 변환이 가능한 시스템을 구현한다. 우선, 입력 영상에 대해 변환된 영상을 출력하는 generative network 를 정의한다. 또한, 변환된 영상을 다시 본래 영상으로 변환하는 inverse network 를 정의한다. Generative network 와 inverse network 를 모두 통과한 결과 영상과 본래 영상의 차 영상을 통해 손실 함수를 정의함으로써 파라미터를 목적에 맞게 학습시킬 수 있다. 또한, generative network 를 통과한 결과 영상과 목적하는 영상을 구분하는 discrimination network 를 정의하여 discrimination network 와 generative network 의 minimax two- player game 을 통해 변환된 영상이 실제 목적 영상과 유사하도록 유도한다. 제안하는 알고리즘을 적용하여 야간 도로 영상의 보정을 수행하면 주변 물체 인식이 어려운 야간 영상을 물체 인식이 용이한 주간 영상으로 변환 할 수 있다.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.16 no.2
• /
• pp.307-312
• /
• 2021

Artificial Intelligence has a long history and is used in various fields including image recognition and automatic translation. Therefore, when we first encounter artificial intelligence, many terms, concepts and technologies often have difficulty in setting or implementing research direction. This study summarized important concepts related to artificial intelligence and summarized the progress of the past 60 years to help researcher suffering from these difficulties. Through this, it is possible to establish the basis for the use of vast artificial intelligence technologies and establish the right direction for research.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.6
• /
• pp.2115-2127
• /
• 2021

Although significant progress has been made in synthesizing visually realistic face images by Generative Adversarial Networks (GANs), there still lacks effective approaches to provide fine-grained control over the generation process for semantic facial attribute editing. In this work, we propose a novel cross channel self-attention based generative adversarial network (CCA-GAN), which weights the importance of multiple channels of features and archives pixel-level feature alignment and conversion, to reduce the impact on irrelevant attributes while editing the target attributes. Evaluation results show that CCA-GAN outperforms state-of-the-art models on the CelebA dataset, reducing Fréchet Inception Distance (FID) and Kernel Inception Distance (KID) by 15~28% and 25~100%, respectively. Furthermore, visualization of generated samples confirms the effect of disentanglement of the proposed model.

• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10a
• /
• pp.129-132
• /
• 1996

There is a tendency nowadays to produce increasingly miniaturized electronic equipment which incorporate parts that have to be precisely positioned, like lenses, heads and CCD's in scanners, printers, copiers, VCR's, optical fiber modules, etc. In contrast to the production process of precision parts, which is currently being carried out automatically, the assemblage process is still being performed by specially skilled technicians. The assemblage process comprises normally the following steps: firstly, the parts are roughly positioned and partially fixed, secondly, the parts are manually nudged towards the target position and finally glued, screwed or welded. This paper presents a system that uses six piezo Impact Drive Mechanisms for accurate micro positioning within three degrees of freedom (lateral and longitudinal translation and rotation). The system is designed to positioning a printed circuit board with an accuracy better than 3 .mu.m (for translations), 5 mrad (for rotation).

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2023.05a
• /
• pp.285-286
• /
• 2023

The visualization of information serves as a valuable tool for facilitating communication and exchange of opinions among stakeholders by conveying information in an intuitive and clear manner. As a preliminary study of visualization for construction field, this study proposed a model for generating three-dimensional floor layout using 360-degree panoramic cameras. The model integrates the layouts by calculating normal vectors of the plane which has openings, and applying translation and rotation matrices between the normal vectors. The results of this study can contribute to improving communication in construction sites by incorporating visualization, and further to the digital transformation of the construction industry.

• Journal of Korean Society of Archives and Records Management
• /
• v.22 no.4
• /
• pp.129-149
• /
• 2022

Records have temporal characteristics, including the past and present; linguistic characteristics not limited to a specific language; and various types categorized in a complex way. Processing records such as text, video, and audio in the life cycle of records' creation, preservation, and utilization entails exhaustive effort and cost. Primary natural language processing (NLP) technologies, such as machine translation, document summarization, named-entity recognition, and image recognition, can be widely applied to electronic records and analog digitization. In particular, Korean deep learning-based NLP technologies effectively recognize various record types and generate record management metadata. This paper provides an overview of Korean NLP technologies and discusses considerations for applying NLP technology in records management. The process of using NLP technologies, such as machine translation and optical character recognition for digital conversion of records, is introduced as an example implemented in the Python environment. In contrast, a plan to improve environmental factors and record digitization guidelines for applying NLP technology in the records management field is proposed for utilizing NLP technology.

• Korean Journal of Remote Sensing
• /
• v.27 no.2
• /
• pp.141-150
• /
• 2011

Precise image-to-image registration is required to process multi-sensor data together. The purpose of this paper is to develop an algorithm that register between high-resolution optical and SAR images using linear features. As a pre-processing step, initial alignment was fulfilled using manually selected tie points to remove any dislocations caused by scale difference, rotation, and translation of images. Canny edge operator was applied to both images to extract linear features. These features were used to design a cost function that finds matching points based on their similarity. Outliers having larger geometric differences than general matching points were eliminated. The remaining points were used to construct a new transformation model, which was combined the piecewise linear function with the global affine transformation, and applied to increase the accuracy of geometric correction.