• International Journal of Automotive Technology
• /
• 제6권1호
• /
• pp.53-63
• /
• 2005

We propose a fuzzy neural network (FNN) theory capable of deciding the quality of a road image prior to extracting lane-related information. The accuracy of lane-related information obtained by image processing depends on the quality of the raw images, which can be classified as good or bad according to how visible the lane marks on the images are. Enhancing the accuracy of the information by an image-processing algorithm is limited due to noise corruption which makes image processing difficult. The FNN, on the other hand, decides whether road images are good or bad with respect to the degree of noise corruption. A cumulative distribution function (CDF), a function of edge histogram, is utilized to extract input parameters from the FNN according to the fact that the shape of the CDF is deeply correlated to the road image quality. A suitability analysis shows that this deep correlation exists between the parameters and the image quality. The input pattern vector of the FNN consists of nine parameters in which eight parameters are from the CDF and one is from the intensity distribution of raw images. Experimental results showed that the proposed FNN system was quite successful. We carried out simulations with real images taken in various lighting and weather conditions, and obtained successful decision-making about $99\%$ of the time.

• 지능정보연구
• /
• 제8권1호
• /
• pp.27-39
• /
• 2002

본 논문은 다양한 웹 이미지로부터 문자영역(text block)의 위치를 알아내고 문자영역을 추출하는 방법을 제안한다. 인터넷 사용자관점에서 볼 때, 웹 이미지에 포함되어 있는 문자정보는 중요한 정보이지만 최근까지 이 분야의 연구는 그리 활발하지 못했다. 본 연구에서 제안된 알고리즘은 문자의 경사방향(skew)과 문자의 크기나 폰트에 관한 사전 정보 없이 수행되어 질 수 있도록 제안되었다 폰트 스타일과 크기에 제약되지 않고 문자영역을 적합하게 추출하기 위해 유용한 에지 검출, 문자 클러스터링 영역으로 정의되는 문자의 고유한 특성을 위한 히스토그램을 사용하였다. 다수의 실험을 통하여 제안된 방법을 테스트하고 수용할 만한 결과를 도출했다.

• 한국정보처리학회논문지
• /
• 제7권9호
• /
• pp.2994-3001
• /
• 2000

본 논문에서는 영역병합 방법을 이용한 칼라 영상 분할 방법을 제안하였다. 영상 분할 전단계에서 비선형 필터링 방법을 이용한 평활화와 채도 강화 및 명도 평균화를 수행하여, 영상 내 존재하는 비균질성을 줄이고, 칼라 히스토그램의 zero-crossing 정보를 이용한 비균일 양자화를 수행하여 유사한 칼라성분을 가지는 영역들을 분할하였다. 웨이브릿 변환의 고주파 대역 에너지를 이용하여 분할된 초기 영역의 윤곽성분 강도를 측정하였고, 이를 통해 병합 후 후보영역을 선정하였다. 영역병합을 위한 영역간 유사도 측정은 R, G, B 칼라성분의 유클리디안 거리를 측정하여 수행하였다. 제안된 방법은 기존의 방법에 비해 불규칙한 광원으로 불필요한 영역이 분할되는 것을 줄일 수 있었고, 이를 실험을 통해 입증하였다.

• 한국멀티미디어학회논문지
• /
• 제7권12호
• /
• pp.1657-1664
• /
• 2004

We propose a method that classifies images into two object types man-made and natural objects. A central object is extracted from each image by using central object extraction method[1] before classification. A central object in an images defined as a set of regions that lies around center of the image and has significant color distribution against its surrounding. We define three measures to classify the object images. The first measure is energy of edge direction histogram. The energy is calculated based on the direction of only non-circular edges. The second measure is an energy difference along directions in Gabor filter dictionary. Maximum and minimum energy along directions in Gabor filter dictionary are selected and the energy difference is computed as the ratio of the maximum to the minimum value. The last one is a shape of an object, which is also represented by Gabor filter dictionary. Gabor filter dictionary for the shape of an object differs from the one for the texture in an object in which the former is computed from a binarized object image. Each measure is combined by using majority rule tin which decisions are made by the majority. A test with 600 images shows a classification accuracy of 86%.

• 한국통신학회논문지
• /
• 제30권11C호
• /
• pp.1076-1082
• /
• 2005

최근 수리형태학적 접근 방법을 이용하여 영상을 분할하고자 하는 연구가 계속되고 있다. 그 중에서도 분수경계 알고리듬은 기존의 에지 기반의 영상 분할 방법과 영역기반의 영상분할 방법의 장점을 모두 가지고 있는 효과적인 영상 분할 기법 중에 하나이다. 분수경계 알고리듬의 기본적인 개념은 지형학적 해석에 기반을 두고 있으며 항상 영역의 외곽에 폐곡선을 형성한다. 그러나 잡영에 매우 민감하게 반응하여 수많은 영역으로 분할되는 과분할 현상을 초래한다. 따라서 본 논문에서는 중요하지 많은 국부 최소점과 국부 최대점을 모두 제거함으로써 과분할 현상을 줄이는 제한적 워터폴 알고리듬을 제안한다. 실험결과 제안한 제한적 워터폴 방법이 다른 과분할 억제 방법보다 평균분할 영역수와 외곽선 소실 측면에서 효과적으로 영상을 분할할 수 있었다.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2005년도 ICCAS
• /
• pp.1287-1292
• /
• 2005

This paper presents an effective pattern classification model by designing an artificial neural network based pattern classifiers for face recognition. First, a RGB image inputted from a frame grabber is converted into a HSV image which is similar to the human beings' vision system. Then, the coarse facial region is extracted using the hue(H) and saturation(S) components except intensity(V) component which is sensitive to the environmental illumination. Next, the fine facial region extraction process is performed by matching with the edge and gray based templates. To make a light-invariant and qualified facial image, histogram equalization and intensity compensation processing using illumination plane are performed. The finally extracted and enhanced facial images are used for training the pattern classification models. The proposed H-ART2 model which has the hierarchical ART2 layers and F-LVQ model which is optimized by fuzzy membership make it possible to classify facial patterns by optimizing relations of clusters and searching clustered reference patterns effectively. Experimental results show that the proposed face recognition system is as good as the SVM model which is famous for face recognition field in recognition rate and even better in classification speed. Moreover high recognition rate could be acquired by combining the proposed neural classification models.

• 제어로봇시스템학회논문지
• /
• 제8권6호
• /
• pp.468-476
• /
• 2002

The confidence of information from image processing depends on the original image quality. Enhancing the confidence by an algorithm has an essential limitation. Especially, road images are exposed to lots of noisy sources, which makes image processing difficult. We, in this paper, propose a FNN (fuzzy-neural network) capable oi deciding the quality of a road image prior to extracting lane-related information. According to the decision by the FNN, road images are classified into good or bad to extract lane-related information. A CDF (cumulative distribution function), a function of edge histogram, is utilized to construct input parameters of the FNN, it is based on the fact that the shape of the CDF and the image quality has large correlation. Input pattern vector to the FNN consists of ten parameters in which nine parameters are from the CDF and the other one is from intensity distribution of raw image. Correlation analysis shows that each parameter represents the image quality well. According to the experimental results, the proposed FNN system was quite successful. We carried out simulations with real images taken by various lighting and weather conditions and achieved about 99% successful decision-making.

• 제어로봇시스템학회논문지
• /
• 제21권3호
• /
• pp.276-282
• /
• 2015

This paper proposes a SMD (surface mounting device) classification method for the PCB assembly inspection machines. The package types of SMD components should be classified to create the job program of the inspection machine. In order to reduce the creation time of job program, we developed the automatic classification algorithm for the SMD packages. We identified the chip-type packages by color and edge distribution of the images. The input images are transformed into the HSI color model, and the binarized histroms are extracted for H and S spaces. Also the edges are extracted from the binarized image, and quantized histograms are obtained for horizontal and vertical direction. The neural network is then applied to classify the package types from the histogram inputs. The experimental results are presented to verify the usefulness of the proposed method.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 1999년도 하계종합학술대회 논문집
• /
• pp.583-586
• /
• 1999

In this paper, the wavelet transform is performed in the input 256$\times$256 color image and decomposes a image into low-pass and high-pass components. Since the high-pass band contains the components of three directions, edges are detected by combining three parts. After finding the position of face using the histogram of the edge component, a face region in low-pass band is cut off. Since RGB color image is sensitively affected by luminances, the image of low pass component is normalized, and a facial region is detected using face color informations. As the wavelet transform decomposes the detected face region into three layer, the dimension of input image is reduced. In this paper, we use the 3000 images of 10 persons, and KL transform is applied in order to classify face vectors effectively. FCM(Fuzzy C-Means) algorithm classifies face vectors with similar features into the same cluster. In this case, the number of cluster is equal to that of person, and the mean vector of each cluster is used as a codebook. We verify the system performance of the proposed algorithm by the experiments. The recognition rates of learning images and testing image is computed using correlation coefficient and Euclidean distance.

• 한국멀티미디어학회논문지
• /
• 제15권7호
• /
• pp.837-846
• /
• 2012

본 논문에서는 안전한 웹 맵 매쉬업 서비스 구현을 위한 위성영상의 원근거리 기반의 위성영상 워터마킹 기법을 제안한다. 제안한 방법에서는 위성 영상의 위치 정보와 사용자 정보를 위성 영상의 원근거리에 따라 에지 및 색상 히스토그램 분포에 은닉한다. 따라서 사용자 요구에 의해 서비스되는 임의 해상도의 위성영상들에 대해 해당 해상도의 영상 특성에 적합한 워터마킹 기법을 수행하여 해당위성 영상을 악의적으로 수정하여 불법 유통한 유통자를 추적하고, 사용자의 사생활 보호 할 수 있다. 실험 결과 제안한 기법의 비가시성이 우수함을 확인하고, 회전, 이동 등의 기하학적 공격에도 삽입된 워터마크가 강인함을 확인한다.