• 한국정밀공학회지
• /
• 제32권12호
• /
• pp.1039-1047
• /
• 2015

Contact type sensors (e.g., displacement sensor and strain gauge) were typically used to evaluate the safety and mechanical properties in machines and construction. However, those contact type sensors have been constrained because of measurement problems such as surface roughness, temperature, humidity, and shape. The Digital Image Correlation (DIC) measurement system is a vision measurement system. This measurement system uses the taken image using a CCD camera and calculates the image correlation between the reference image and the deformed image under external force to measure the displacement and strain rates. In this paper, we discuss methods to improve the measurement precision of the digital image correlation measurement system. A tensile test was conducted to compare the precision improvement effects, by using the universal test machine and the DIC measurement system, with the use of subpixel algorithms, i.e., the Coarse Fine Search (CFS) algorithm and the Peak Finding (PF) algorithm.

• Nuclear Engineering and Technology
• /
• 제52권10호
• /
• pp.2328-2333
• /
• 2020

The filter technique was applied to noise images, as noise is the significant factor that cause poor image quality due to lower photon counting. The purpose of this study is to confirm that image quality can be improved using the median modified Wiener filter (MMWF) technique; this is achieved via a National Electrical Manufacturers Association International Electrotechnical Commission body phantom with four large spheres that are filled with the ^99mTc radioisotope when evaluating the image quality. Conventional filters such as Wiener, Gaussian, and median filters were designed, and signal to noise ratio, coefficient of variation, and contrast to noise ratio were used as the evaluation parameters. The improvement in the image quality was in the following order, from the least to the highest improvement, in all cases: Wiener filter, Gaussian filter, median filter, and the MMWF technique. The results show that the image quality was improved from 20.6 to 65.5%, 7.4-40.3%, and 12.7-44.7% for the SNR, COV, and CNR values, respectively, when using the MMWF technique, compared with the use of conventional filters. In conclusion, our results demonstrated that the MMWF technique is useful for reducing the noise distribution in gamma camera images.

• 전자공학회논문지SC
• /
• 제45권1호
• /
• pp.15-24
• /
• 2008

본 논문은 영상의 블록기반 지역 명암대비 향상을 통한 전역 명암대비 향상 기법을 제안한다. 전역 명암대비 향상 기법은 영상의 특성을 고려하지 않고 히스토그램 평활화만을 통해 명암값의 재분배를 수행함으로써 과도한 밝기값의 화소들을 갖는 결과영상을 유발한다. 반면에 블록기반 지역 명암대비 향상 기법은 블록화 현상 및 이를 줄이는 과정에서 영상의 중요한 특징을 훼손하는 문제점을 갖고 있다. 이런 문제점들을 해결하기 위해 본 논문은 입력영상을 다양한 블록 크기로 분할한 임시영상들에 대해 블록기반 히스토그램 평활화를 수행한다. 그리고 지역 명암대비 향상을 통해 얻어진 전역 히스토그램 평활화 함수를 원 영상에 적용하는 전역 명암대비 향상 기법을 수행한다. 지역 명암대비가 향상된 임시 영상으로부터 전역 히스토그램 평활화 함수를 얻기 때문에 제안하는 방법은 지역 명암대비 향상 기법과 전역 명암대비 향상 기법의 장점들을 갖는다.

• 사물인터넷융복합논문지
• /
• 제9권1호
• /
• pp.131-138
• /
• 2023

본 논문은 RAPGAN(Relativistic Average Patch GAN)과 RRDB(Residual in Residual Dense Block)을 이용한 Image-to-Image 변환의 성능 개선에 관한 것이다. 본 논문은 Image-to-Image 변환의 일종인 기존의 pix2pix의 결점을 보완하기 위해 세 가지 측면의 기술적 개선을 통한 성능 향상을 도모함에 그 목적이 있다. 첫째, 기존의 pix2pix 생성자와 달리 입력 이미지를 인코딩하는 부분에서 RRDB를 이용함으로써 더욱 더 깊은 학습을 가능하게 한다. 둘째, RAPGAN 기반의 손실함수를 사용해 원본 이미지가 생성된 이미지에 비해 얼마나 진짜 같은지를 예측하기 때문에 이 두 이미지가 모두 적대적 생성 학습에 영향을 미치게 된다. 마지막으로, 생성자를 사전학습시켜 판별자가 조기에 학습되는 것을 억제하도록 조치한다. 제안된 방법에 따르면, FID 측면에서 기존의 pix2pix보다 평균 13% 이상의 우수한 이미지를 생성할 수 있었다.

• KIEAE Journal
• /
• 제10권1호
• /
• pp.73-83
• /
• 2010

In urban regeneration, resident participation and respect of residents' need have become a major important issue. The purpose of this study is to identify residential area improvement characteristics expected by local residents. Data used for this study were collected from survey which used questionnaire and village image map construction tool kit, developed for facilitating the residents' participation in an actual housing improvement area at Kwngjoo, Korea. The major contents surveyed through questionnaire were first, future images of the area second, directions of improvement third, preferred architectural types such as high or low rise buildings. and a total of 335 data was collected within 4 days during 12-14 December, 2008. The kit was used by parents of students at a local elementary school, and 205 image maps were collected. Content analysis was to analyse characteristics of villages shown in the constructed image maps. Lynch's five elements were utilized to select areas for analysis. As a result, types of buildings desired by residents at the selected four local areas were identified. In general, residents desired their village to be improved with low and mid rise buildings, respecting existing cultural assets. This study showed that there is certain characteristics in relation to the selected areas. Besides, the tool kit used this study showed the effectiveness in collecting opinions from young households in the improvement area within a short time. The tool is expected to be useful in attracting residents and in facilitating participation of wide range of local residents by improving the constraints stemming from time and space.

• 한국콘텐츠학회논문지
• /
• 제21권8호
• /
• pp.1-9
• /
• 2021

딥 러닝을 활용한 컴퓨터 비전 연구는 여전히 대규모의 학습 데이터와 컴퓨팅 파워가 필수적이며, 최적의 네트워크 구조를 도출하기 위해 많은 시행착오가 수반된다. 본 연구에서는 네트워크 최적화나 데이터를 보강하는 것과 무관하게 데이터 자체의 특성만을 고려한 CR(Confusion Rate)기반의 유사 이미지 분류 성능 향상 기법을 제안한다. 제안 방법은 유사한 이미지 데이터를 정확히 분류하기 위해 CR을 산출하고 이를 손실 함수의 가중치에 반영함으로서 딥 러닝 모델의 성능을 향상시키는 기법을 제안한다. 제안 방법은 네트워크 최적화 결과와 독립적으로 이미지 분류 성능의 향상을 가져올 수 있으며, 클래스 간의 유사성을 고려해 유사도가 높은 이미지 식별에 적합하다. 제안 방법의 평가결과 HanDB에서는 0.22%, Animal-10N에서는 3.38%의 성능향상을 보였다. 제안한 방법은 다양한 Noisy Labeled 데이터를 활용한 인공지능 연구에 기반이 될 것을 기대한다.

• 한국산학기술학회논문지
• /
• 제15권8호
• /
• pp.5207-5212
• /
• 2014

본 연구는 지역축제의 성공적인 운영을 통하여 도시이미지를 개선하고 이를 통해 사회문화적, 경제적 파급효과를 높이기 위한 목적으로 수행되었다. 국내의 많은 지자체에서 운영하고 있는 지역축제의 현실은 경제적 파급효과의 극대화를 도모하는데 포커스가 맞춰져 있다. 그러나, 지역축제는 도시를 중심으로 수행되어야 하며 이렇게 운영될 때 도시가 갖는 이미지가 정립될 수 있고, 장기적으로 다양한 프로그램을 통하여 도시마케팅으로 연계가 가능해진다. 지역으로의 관광객을 흡입하기 위해서 본 연구에서는 지역축제 참가자의 도시이미지 변화에 대한 내용을 중점적으로 다루었으며, 그 개선방안으로서 IT기술을 활용한 모바일 환경조성과 실시간 서비스를 제공함으로써 지역이 갖는 변화와 혁신, 전통과 문화의 조화를 통하여 도시만족도를 제고할 수 있는 대안을 제시하고자 하였다.

• 한국실내디자인학회논문집
• /
• 제17권5호
• /
• pp.91-98
• /
• 2008

The cafeteria is one of the welfare facilities in the college and crowed with students during lunchtime because of the various menu, saying time and cheap price. In designing the cafeteria, we have to consider the users' needs and the factors which effect on their satisfaction and behavior under the pleasant environment. The purpose of this study is to draw the environmental improvement factors by Post Occupancy Evaluation. A survey about the users' actual condition and environmental satisfaction level and a behavior observation in the field were utilized as the study methode. The 150 copies of questionnaire were completed by students who ate at the cafeteria of W college in Daejeon city. The SPSS 15.0 for Windows was used for a statistical analysis. The results are listed below; two factors were drawn as the improvement elements through the combination of the unsatisfactory items obtained by the descriptive statistics and the influential items by the multi-regression analysis. They are images and approach. In the 'image' category, there us a statistical difference between the present image and prefer image. The users prefer much brighter, warmer, newer, more gorgeous, more cheerful, more clear, more active, more comfortable, and more modern image of cafeteria. In the 'approach', the location of ail existing cafeteria is inaccessible from the classrooms in spite of an important relationship between classrooms and cafeteria. Therefore the complementary plan for convenient approach is demanded for improvement of users' satisfaction.

• International Journal of Computer Science & Network Security
• /
• 제24권1호
• /
• pp.9-16
• /
• 2024

AI or Artificial Intelligence has been a significant tool used in the organisational backgrounds for an effective improvement in the management methods. The processing of the information and the analysis of the data for the further achievement of heightened efficiency can be performed by AI through its data analytics measures. In the medical field, AI has been integrated for an improvement within the management of the medical services and to note a rise in the levels of customer satisfaction. With the benefits of reasoning and problem solving, AI has been able to initiate a range of benefits for both the consumers and the medical personnel. The main benefits which have been noted in the integration of AI would be integrated into the study. The issues which are noted with the integrated AI usage for the medical sector would also be identified in the study. Medical Image Processing has been seen to integrate 3D image datasets with the medical industry, in terms of Computed Tomography (CT) or Magnetic Resonance Imaging (MRI). The usage of such medical devices have occurred in the diagnosis of the patients, the development of guidance towards medical intervention and an overall increase in the medical efficiency. The study would focus on such different tools, adhered with AI for increased medical improvement.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2001년도 제14회 신호처리 합동 학술대회 논문집
• /
• pp.669-672
• /
• 2001

We propose an efficient algorithm for an interactive content-based image retrieval using relevance feedback. In the proposed algorithm, a new query feature vector first is yielded from the average feature vector of the relevant images that is fed back from the result images of the previous retrieval. Each component weight of a feature vector is computed from an inverse of standard deviation for each component of the relevant images. The updated feature vector of the query and the component weights are used in the iterative retrieval process. In addition, the irrelevant images are excluded from object images in the next iteration to obtain additional performance improvement. In order to evaluate the retrieval performance of the proposed method, we experiment for three image databases, that is, Corel, Vistex, and Ultra databases. We have chosen wavelet moments, BDIP and BVLC, and MFS as features representing the visual content of an image. The experimental results show that the proposed method yields large precision improvement.