• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.7 no.4
• /
• pp.839-844
• /
• 2003

To improve classification accuracy in this paper, we proposed an algorithm for the chromosome image reconstruction in the image preprocessing part. also we proposed the pattern classification method using the hierarchical multilayer neural network(HMNN) to classify the chromosome karyotype. It reconstructed chromosome images for twenty normal human chromosome by the image reconstruction algorithm. The four morphological and ten density feature parameters were extracted from the 920 reconstructed chromosome images. The each combined feature parameters of ten human chromosome images were used to learn HMNN(Hierarchical Multilayer Neural Network) and the rest of them were used to classify the chromosome images. The experimental results in this paper were composed to optimized HMNN and also obtained about 98.26% to recognition ratio.

• The KIPS Transactions:PartB
• /
• v.15B no.5
• /
• pp.435-448
• /
• 2008

3D posture recognition is a solution to overcome the limitation of 2D posture recognition. There are many researches carried out for 3D posture recognition using 3D data. The 3D data consist of massive surface points which are rich of information. However, it is difficult to extract the important features for posture recognition purpose. Meanwhile, it also consumes lots of processing time. In this paper, we introduced a dimension reduction method that transform 3D surface points of an object to 2D data representation in order to overcome the issues of feature extraction and time complexity of 3D posture recognition. For a better feature extraction and matching process, a cylindrical boundary is introduced in meshless parameterization, its offer a fast processing speed of dimension reduction process and the output result is applicable for recognition purpose. The proposed approach is applied to hand and human posture recognition in order to verify the efficiency of the feature extraction.

• The KIPS Transactions:PartB
• /
• v.13B no.1 s.104
• /
• pp.47-52
• /
• 2006

This paper proposes a method of content-based image retrieval using region feature vector in order to overcome disadvantages of existing color histogram methods. The color histogram methods have a weak point that reduces accuracy because of quantization error, and more. In order to solve this, we convert color information to HSV space and quantize hue factor being purecolor information and calculate histogram and then use thus for retrieval feature that is robust in brightness, movement, and rotation. Also we solve an insufficient part that is the most serious problem in color histogram methods by dividing an image into sixteen regions and then comparing each region. We improve accuracy by edge and DC of DCT transformation. As a result of experimenting with 1,000 color images, the proposed method has showed better precision than the existing methods.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.12 no.1
• /
• pp.13-21
• /
• 2019

The content based on image retrieval makes use of features of information within image such as color, texture and share for Retrieval data. we present a novel approach for improving retrieval accuracy based on DCT Filter-Bank. First, we perform DCT on a given image, and generate a Filter-Bank using the DCT coefficients for each color channel. In this step, DC and the limited number of AC coefficients are used. Next, a feature vector is obtained from the histogram of the quantized DC coefficients. Then, AC coefficients in the Filter-Bank are separated into three main groups indicating horizontal, vertical, and diagonal edge directions, respectively, according to their spatial-frequency properties. Each directional group creates its histogram after employing Otsu binarization technique. Finally, we project each histogram on the horizontal and vertical axes, and generate a feature vector for each group. The computed DC and AC feature vectors bins are concatenated, and it is used in the similarity checking procedure. We experimented using 1,000 databases, and as a result, this approach outperformed the old retrieval method which used color information.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2015.07a
• /
• pp.177-180
• /
• 2015

본 논문에서는 비디오 영상을 입력 했을 때 특징점 추적을 통한 다수 영상의 고속 스트칭 기법을 제안한다. 빠른 속도로 특징점 추출을 위해서 FAST(Features from Accelerated Segment Test) 기법을 사용한다. 특징점 정합과정은 기존의 방법과는 다른 새로운 방법을 제안한다. Mean shift 를 통해 특징점이 포함된 영역을 추적하여 벡터(vector)를 구한다. 이 벡터를 사용하여 추출한 특징점들을 정합하는데 사용한다. 마지막으로 이상점(outlier)을 제거하기 위해 RANSAC(RANdom Sample Consensus) 기법을 사용한다. 입력된 두 영상의 호모그래피(homography) 변환 행렬을 구하여 하나의 파노라마 영상을 생성한다. 실험을 통해 제안하는 기법이 기존의 기법보다 속도가 향상되는 것을 확인하였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2010.06c
• /
• pp.390-394
• /
• 2010

본 논문에서는 영상인식에서 널리 사용되는 지역적 특징인 SIFT와 부분공간분석에 의한 차원축소방법의 결합을 통하여 얼굴을 인식하는 방법을 제안한다. 기존의 SIFT기반 영상인식 방법에서는 추출된 키 포인트 각각에 대하여 계산된 특징기술자들을 개별적으로 비교하여 얻어지는 유사도를 바탕으로 인식을 수행하는데 반해, 본 논문에서 제안하는 접근법은 SIFT의 특징기술자를 명도 값으로 표현된 얼굴 영상을 여려 변형에 강건한 형태로 표현되도록 변환하는 표현방식으로 본다. SIFT기반의 특징기술자에 의해 표현된 얼굴 영상을 부분공간분석법에 의해 저차원의 특징벡터로 다시 표현되고, 이 특징벡터를 이용하여 얼굴인식을 수행한다. 잘 알려진 벤치마크 데이터인 AR 데이터베이스에 대한 실험을 통해 제안한 방법이 조명 변화와 가려짐에 강인한 인식 결과를 보여줄 뿐 아니라, 기존의 SIFT 기반의 얼굴 인식 방법에 비하여 우수한 처리 속도를 보임을 확인하였다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2018.11a
• /
• pp.131-132
• /
• 2018

비디오 스티칭은 카메라 간 변환 관계인 호모그래피를 이용하여 스티칭하는 것이 일반적이다. 본 논문은 호모그래피를 이용한 고정형 비디오 스티칭에서 조도 변화, 노이즈 등으로 일관되지 않는 특징점 추출과 유니폼한 입력 영상으로 적은 특징점이 추출되는 경우에 대하여 정확도 높은 호모그래피 추출이 가능한 특징점 누적 기반 고정형 비디오 스티칭 방법을 제안한다. 실험을 통해 단일 프레임 특징점을 이용한 결과 영상에 비해 특징점 누적을 이용하는 경우 영상 내 부정합 영역 등의 왜곡이 크게 감소하였음을 확인하였다.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.42 no.2 s.302
• /
• pp.11-18
• /
• 2005

To detect vehicles in image, first the image is transformed with the steerable pyramid which has independent directions and levels. Feature vectors are the collection of filter responses at different scales of a steerable image pyramid. For the detection of vehicles in image, feature vectors in feature points of the vehicle image is used. First the feature points are selected with the grid points in vehicle image that are evenly spaced, and second, the feature points are comer points which m selected by human, and last the feature points are corner Points which are selected in grid points. Next the feature vectors of the model vehicle image we compared the patch of the test images, and if the distance of the model and the patch of the test images is lower than the predefined threshold, the input patch is decided to a vehicle. In experiment, the total 11,191 vehicle images are captured at day(10,576) and night(624) in the two local roads. And the $92.0\%$ at day and $87.3\%$ at night detection rate is achieved.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.24 no.6
• /
• pp.590-596
• /
• 2004

Neural network-based signal classification systems are increasingly used in the analysis of large volumes of data obtained in NDE applications. Ultrasonic inspection methods on the other hand are commonly used in the nondestructive evaluation of welds to detect flaws. An important characteristic of ultrasonic inspection is the ability to identify the type of discontinuity that gives rise to a peculiar signal. Standard techniques rely on differences in individual A-scans to classify the signals. This paper proposes an ultrasonic signal classification technique based on the information tying in the neighboring signals. The approach is based on a 2-dimensional Fourier transform and the principal component analysis to generate a reduced dimensional feature vector for classification. Results of applying the technique to data obtained from the inspection of actual steel welds are presented.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2013.10a
• /
• pp.103-107
• /
• 2013

In this paper, A real-time lane tracking algorithm is proposed for lane departure warning system. To eliminate perspective effect, input image is converted into Bird's View by inverse perspective mapping. Next, suitable features are extracted for lane detection. Lane feature that correspond to area of interest and RANSAC are used to detect lane candidates. And driving lane is decided by clustering of lane candidates. Finally, detected lane is tracked using the Kalman filter. Experimental results show that the proposed algorithm can be processed within 30ms and its detection rate is approximately 90% on the highway in a variety of environments such as day and night.