• 전기전자학회논문지
• /
• 제27권1호
• /
• pp.59-64
• /
• 2023

딥러닝은 다양한 컴퓨터 비전 문제를 해결할 수 있지만, 대량의 데이터셋이 필요하다. 본 논문에서는 대형 이미지 데이터셋을 구축하기 위해 이미지 이진화 기반 데이터 증강 기법을 제안한다. 이미지 이진화를 사용하여 특성을 추출하고 추출된 나머지 픽셀을 랜덤하게 배치하여 새로운 이미지를 생성한다. 생성된 이미지는 원본 이미지와 유사한 품질을 보여주며, 딥러닝 모델에서도 뛰어난 성능을 보였다.

• 한국산업정보학회:학술대회논문집
• /
• 한국산업정보학회 2001년도 춘계학술대회논문집:21세기 신지식정보의 창출
• /
• pp.64-76
• /
• 2001

In this paper, A simple image hologram encryption and decryption technique based on the principle of interference are proposed. The technique using the photorefractive material for getting a stable interference pattern is also proposed. And combine these two techniques, I would like to implement a stable optical information security system. In the encrypting process, I would generate binary phase hologram which can reconstruct original image perfectly, and regard this hologram as original image to be encrypted image. And then the hologram is encrypted as randomly generated binary phase image. Reference image is also generated from the encrypted image by applying interference rule. In the decrypting process, I can get a interference intensity by interfering the reference image and the encrypted image in the interferometer. and transform inferference intensity information into phase information. I recover original image by inverse Fourier transforming the phase information. In this process, the intensity information generated by interference of two images is very sensitive to external vibrations. So, I would like to get a stable interference using the characteristic of SPPCM(self pumped phase conjugate mirror) in photorefractive materials, especially BaTiO₃.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2005년도 Proceedings of ISRS 2005
• /
• pp.571-574
• /
• 2005

In this paper, we propose an efficient image segmentation technique for large-sized monochrome images using a hybrid approach which combines threshold and region-based techniques. First, an image is partitioned into fixed-size blocks and for each block the representative intensity is determined by averaging pixel intensities within the block. Next, the neighborhood blocks that have similar characteristics with respect to a specific threshold are merged in order to form candidate regions. Finally, those candidate regions are refined to get final target object regions by merging regions considering the spatial locality and certain criteria. We have performed experiments on images selected from various domains and showed that our technique was able to extract target object regions appropriately from most images.

• 한국인터넷방송통신학회논문지
• /
• 제14권4호
• /
• pp.29-37
• /
• 2014

다시점 비디오는 카메라 수만큼의 컬러영상들과 깊이 영상들이 존재하므로 막대한 양의 데이터를 갖게 되어서, 그 양을 줄이는 새로운 압축 기술 개발이 필수적이다. 최근 계층적 깊이 영상(Layered Depth Image : LDI)의 개념을 이용하여 깊이 정보를 포함하는 다시점 비디오를 효과적으로 압축 부호화하는 기술이 주목 받고 있다. 이 방법은 여러 시점의 깊이 정보와 워핑(warping) 함수를 사용하여 다시점의 컬러와 깊이 영상을 합성하여 하나의 데이터 구조를 만드는 방법이다. 본 논문에서는 계층적 깊이 영상에서 실제 거리를 이용하고 오버랩(overlap) 문제를 해결하고 그리고 컬러 변환을 기반으로 복원을 위해서 추가되는 데이터의 양을 감소시켜 압축 성능을 향상시키는 방법을 제안하였다.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2006년도 Proceedings of ISRS 2006 PORSEC Volume II
• /
• pp.643-646
• /
• 2006

This paper addresses the approach to extract linear features from satellite imagery using an efficient segmentation method. The extraction of linear features from satellite images has been the main concern of many scientists. There is a need to develop a more capable and cost effective method for the Iranian map revision tasks. The conventional approaches for producing, maintaining, and updating GIS map are time consuming and costly process. Hence, this research is intended to investigate how to obtain linear features from SPOT satellite imagery. This was accomplished using a discontinuity-based segmentation technique that encompasses four stages: low level bottom-up, middle level bottom-up, edge thinning and accuracy assessment. The first step is geometric correction and noise removal using suitable operator. The second step includes choosing the appropriate edge detection method, finding its proper threshold and designing the built-up image. The next step is implementing edge thinning method using mathematical morphology technique. Lastly, the geometric accuracy assessment task for feature extraction as well as an assessment for the built-up result has been carried out. Overall, this approach has been applied successfully for linear feature extraction from SPOT image.

• Nuclear Engineering and Technology
• /
• 제54권10호
• /
• pp.3943-3948
• /
• 2022

Single-photon emission computed tomography is one of the reliable pin-by-pin verification techniques for spent-fuel assemblies. One of the challenges with this technique is to increase the total fuel assembly verification speed while maintaining high verification accuracy. The aim of the present study, therefore, was to develop an artificial intelligence (AI) algorithm-based tomographic image analysis technique for partial-defect verification of fuel assemblies. With the Monte Carlo (MC) simulation technique, a tomographic image dataset consisting of 511 fuel-rod patterns of a 3 × 3 fuel assembly was generated, and with these images, the VGG16, GoogLeNet, and ResNet models were trained. According to an evaluation of these models for different training dataset sizes, the ResNet model showed 100% pattern estimation accuracy. And, based on the different tomographic image qualities, all of the models showed almost 100% pattern estimation accuracy, even for low-quality images with unrecognizable fuel patterns. This study verified that an AI model can be effectively employed for accurate and fast partial-defect verification of fuel assemblies.

• 디지털콘텐츠학회 논문지
• /
• 제2권2호
• /
• pp.157-166
• /
• 2001

본 논문은 내용 기반 이미지 검색 시스템에서 이미지의 위치 및 모양 정보에 의한 회귀선을 추정하여 효율적으로 특징 벡터 추출함과 동시에 같은 도메인상의 특징 벡터가 일정 수준보다 많아질 경우 효율적으로 특징 벡터의 차원을 줄이는 기법을 제안한다. 특히, 특징 벡터의 차원을 줄이는 제안된 기법은 특징 벡터의 수에 관계없이 특정한 n개의 특징 벡터로의 변환이 가능하다. 본 논문에서 제안된 기법들은 실제 내용 기반 이미지 검색 시스템의 구현을 통해 기존의 방법보다 효율적인 검색은 물론 다차원 특징 벡터를 특정 n차원의 특징 벡터로 변환함으로써 다차원 색인 기법이 가지고 있는 가장 큰 단점인 '차원의 저주' 문제를 근본적으로 해결할 수 있는 방법임을 보인다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2001년도 하계종합학술대회 논문집(2)
• /
• pp.63-66
• /
• 2001

In this paper a binary image encryption technique and decryption system based on a joint transform correlator (JTC) are Proposed. In this method, a Fourier transform of the encrypted image is used as the encrypted data and a Fourier transform of the random phase is used as the key code. The original binary image can be reconstructed on a square law device, such as a CCD camera after the joint input is inverse Fourier transformed. The proposed encryption technique does not suffer from strong auto-correlation terms appearing in the output plane. Based on computer simulations, the proposed encryption technique and decoding system were demonstrated as adequate for optical security applications.

• 제어로봇시스템학회논문지
• /
• 제11권2호
• /
• pp.122-126
• /
• 2005

A technique for alignment to true north is presented based on synchronized measurements of vision image by a camera and output voltage of wind direction sensor. The true wind direction is evaluated by means of image processing techniques with least square sense, and then evaluated true value is compared with measured output voltage of the sensor. The uncertainty analysis about the component error for the proposed method in practical situation is performed. The proposed technique is applied to real meteorological tower (wind measuring tower) at the Daekwanryung test site. In addition, some uncertainty analysis of this method is presented.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2003년도 하계종합학술대회 논문집 Ⅳ
• /
• pp.1811-1814
• /
• 2003

In this paper, a novel watermarking technique is proposed to authenticate the owner-ship of copyright for the digital contents. Using the 2-level DWT(Discrete Wavelet Transform) we divide a specific frequency band into detailed blocks and apply PIM(picture information measure). After the complexity is calculated, the watermark is embedded in only on high complexity areas. Conventional watermarking technique damages to the original image, because it does not consider the feature of the whole area or a specific frequency band. Easily affected by noise and compression, it is difficult to extract the watermark. However, the proposed watermarking technique, considering the complexity of input image, does not damage to the original image Simulation result show that the proposed technique has the robustness of JPEG compression, noise and filtering such as a general signal processing