Earticle

In this paper, an accurate segmentation algorithm of video dynamic background image (VDBI) based on improved wavelet transform is proposed. Based on the smooth processing of VDBI, the traditional wavelet transform process is improved, and the two-layer decomposition of dynamic image is realized by using two-dimensional wavelet transform. On the basis of decomposition results and information enhancement processing, image features are detected, feature points are extracted, and quantum ant colony algorithm is adopted to complete accurate segmentation of the image. The maximum SNR of the output results of the proposed algorithm can reach 73.67 dB, the maximum time of the segmentation process is only 7 seconds, the segmentation accuracy shows a trend of decreasing first and then increasing, and the global maximum value can reach 97%, indicating that the proposed algorithm effectively achieves the design expectation.

This paper is an attempt to design segmentation method based on fully convolutional networks (FCN) and attention mechanism. The first five layers of the Visual Geometry Group (VGG) 16 network serve as the coding part in the semantic segmentation network structure with the convolutional layer used to replace pooling to reduce loss of image feature extraction information. The up-sampling and deconvolution unit of the FCN is then used as the decoding part in the semantic segmentation network. In the deconvolution process, the skip structure is used to fuse different levels of information and the attention mechanism is incorporated to reduce accuracy loss. Finally, the segmentation results are obtained through pixel layer classification. The results show that our method outperforms the comparison methods in mean pixel accuracy (MPA) and mean intersection over union (MIOU).

This study proposed a novel framework for plant leaf disease identification. The proposed model consists of four steps including pre-processing, segmentation, feature extraction, and classification. At first, the unwanted noise and overfitting are removed, and also image contrast level is enhanced. Secondly, the Fuzzy C-Means (FCM) based Chameleon Swarm Algorithm (CSA) named as (FCM-CSA) is used for plant leaf diseased part segmentation. In the third stage, the feature extraction is performed using a fast GLCM feature extraction model. Finally, the Progressive Neural Architecture Search (PNAS) is used for plant leaf disease identification. The experimental investigations are carried out using MATLAB software with the Mendeley database. From this dataset, we have used Apple Cedar Apple Rust (ACAR), Cherry Powdery Mildew (CPM), Corn Common Rust (CCCR), Apple Healthy (AH), Grape Black Rot (GBR), Pepper Bell Bacterial Spot (PBBS), Potato Late Blight (PLB) and Tomato Leaf Mold (TLM) disease images. Different measures such as precision, recall, sensitivity, specificity, and accuracy results are used to validate the performance of the proposed model.

In recent years, edge computing technology consists of several Internet of Things (IoT) devices with embedded sensors that have improved significantly for monitoring, detection, and management in an environment where big data is commercialized. The main focus of edge computing is data optimization or task offloading due to data and task-intensive application development. However, existing offloading approaches do not consider correlations and associations between data and tasks involving edge computing. The extent of collaborative offloading segmented without considering the interaction between data and task can lead to data loss and delays when moving from edge to edge. This article proposes a range segmentation of dynamic offloading (RSDO) algorithm that isolates the offload range and collaborative edge node around the edge node function to address the offloading issue.The RSDO algorithm groups highly correlated data and tasks according to the cause of the overload and dynamically distributes offloading ranges according to the state of cooperating nodes. The segmentation improves the overall performance of edge nodes, balances edge computing, and solves data loss and average latency.

Text line segmentation is a critical task in handwritten document recognition. In this paper, we propose a novel text-line-segmentation method using baseline estimation and watershed. The baseline-detection algorithm estimates the baseline using Adaptive Head-Tail Connection (AHTC) on the document. Then, the watershed method segments the line region using the baseline-detection result. Finally, the text lines are separated by watershed result and a post-processing algorithm defines the lines more correctly. The scheme successfully segments text lines with 97% accuracy from the handwritten document images in the ICDAR database.

Level set methods are the numerical techniques for tracking interfaces and shapes. They have been successfully used in image segmentation. A new variational level set evolving algorithm without re-initialization is presented in this paper. It consists of an internal energy term that penalizes deviations of the level set function from a signed distance function, and an external energy term that drives the motion of the zero level set toward the desired image feature. This algorithm can be easily implemented using a simple finite difference scheme. Meanwhile, not only can the initial contour can be shown anywhere in the image, but the interior contours can also be automatically detected.

The image segmentation is the most important operation in an image processing system. It is located at the joint between the processing and analysis of the images. Unsupervised segmentation aims to automatically separate the image into natural clusters. However, because of its complexity several methods have been proposed, specifically methods of optimization. In our work we are interested to the technique SFLA (Shuffled Frog-Leaping Algorithm). It's a memetic meta-heuristic algorithm that is based on frog populations in nature searching for food. This paper proposes a new approach of unsupervised image segmentation based on SFLA method. It is implemented and applied to different types of images. To validate the performances of our approach, we performed experiments which were compared to the method of K-means.

In this paper, we propose a method to build an accurate initial trimap for the GrabCut algorithm without the need for human interaction. First, we identify a rough candidate for the label region of a bottle by applying a saliency map to find a salient area from the image. Then, the Hough Transformation method is used to detect the left and right borders of the label region, and the k-means algorithm is used to localize the upper and lower borders of the label of the bottle. These four borders are used to build an initial trimap for the GrabCut method. Finally, GrabCut segments accurate regions for the label. The experimental results for 130 wine bottle images demonstrated that the saliency map extracted a rough label region with an accuracy of 97.69% while also removing the complex background. The Hough transform and projection method accurately drew the outline of the label from the saliency area, and then the outline was used to build an initial trimap for GrabCut. Finally, the GrabCut algorithm successfully segmented the bottle label with an average accuracy of 92.31%. Therefore, we believe that our method is suitable for product label recognition systems that automatically segment product labels. Although our method achieved encouraging results, it has some limitations in that unreliable results are produced under conditions with varying illumination and reflections. Therefore, we are in the process of developing preprocessing algorithms to improve the proposed method to take into account variations in illumination and reflections.

In this paper we introduce a new circle detection algorithm for occluded on/off pupil and iris boundary extraction. The proposed algorithm employs 7-step processing to detect a center and radius of occluded on/off eye images using the property of the chords. The algorithm deals with two types of occluded pupil and iris boundary information; one is composed of circle-shaped, incomplete objects, which is called occluded on iris images and the other type consists of arc objects in which circular information has partially disappeared, called occluded off iris images. This method shows that the center and radius of iris boundary can be detected from as little as one-third of the occluded on/off iris information image. It is also shown that the proposed algorithm computed the center and radius of the incomplete iris boundary information which has partially occluded and disappeared. Experimental results on RGB images and IR images show that the proposed method has encouraging performance of boundary detection for pupil and iris segmentation. The experimental results show satisfactorily the detection of circle from incomplete circle shape information which is occluded as well as the detection of pupil/iris boundary circle of the occluded on/off image.

The vehicle license plate recognition (VLPR) system analyzes and monitors the speed of vehicles, theft of vehicles, the violation of traffic rules, illegal parking, etc., on the motorway. The VLPR consists of three major parts: license plate detection (LPD), license plate character segmentation (LPCS), and license plate character recognition (LPCR). This paper presents an efficient method for the LPCS and LPCR of Korean vehicle license plates (LPs). LP tilt adjustment is a very important process in LPCS. Radon transformation is used to correct the tilt adjustment of LP. The global threshold segmentation method is used for segmented LP characters from two different types of Korean LPs, which are a single row LP (SRLP) and double row LP (DRLP). The cross-correlation matching method is used for LPCR. Our experimental results show that the proposed methods for LPCS and LPCR can be easily implemented, and they achieved 99.35% and 99.85% segmentation and recognition accuracy rates, respectively for Korean LPs.

Regarding high-dimensional heterogeneous data, combined with the existing algorithms' poor mining accuracy and parameter sensitivity, this paper proposes a local outlier mining algorithm based on neighborhood density. Use region segmentation to split high-dimensional data into reasonable sub-regions, reducing the difficulty of processing a large amount of high-dimensional data. The kernel neighborhood density is used to replace the average neighborhood density, so that the density calculation has nothing to do with data heterogeneity. Finally, the neighborhood state and outlier state of the data are further determined on the basis of neighborhood density to improve the accuracy of outlier mining. Through artificial and UCI data set simulation results, it shows that data volume and data dimension are the main factors that affect data outlier mining. The accuracy, coverage, and efficiency of the algorithm proposed in this paper are significantly better than those of the comparison algorithm, and it has better adaptability to different types of data sets.

영상 세션화는 기본적으로 이미지를 의미있는 구조로 나누는 것이다. 항상 특정 요구에 따라 이미지를 완전히 분할 한다. 하지만 결과가 원하지 않는 오버세션화가 이루어질 수 있다. 따라서 본 논문에서는 오버세션화를 줄이기 위해 가우시안 혼합 모델(GMM) 기반의 변형된 워터쉐드 알고리즘을 제안한다. 거리 변환 영상은 가우시안 혼합 모델에 의해 처리된다. 그리고 기대 최대화(EM) 알고리즘은 가우시안 혼합모델 데이터에 맞추는 데 사용된다. 이들 평균 의 평균을 최적의 임계값으로 선택하고 획득된 임계값에 기초하여 이진 영상으로 변환한다. 이렇게 얻은 이진 영상 은 적절한 구조 요소의 적용하여 침식화한다. 침식된 영상에서 수행된 워터쉐드 분할은 효과적인 분할 결과를 제공 한다. 제안된 방법이 사용될 때 옥수수, 쌀 및 밀과 같은 다른 커널이 효과적으로 분할된다. 실험 목적으로 38 개의 옥수수 커널, 95 개의 쌀 커널 및 32 개의 밀 커널을 사용한다. 분할 정확도는 옥수수, 쌀 및 밀에 대해 각각 94.7 %, 96.8 % 및 90.6 % 이다.

Image segmentation is basically the division of an image into meaningful structures. It always produces complete division of the image as the specific demand. However, it is prone to over-segmentation which makes the result unfavorable. Therefore, this paper proposes modified watershed algorithm based on Gaussian mixture model(GMM) to reduce the over-segmentation. Distance transformed image is processed by GMM. And expectation maximization(EM) algorithm is used to fit GMM to data. Average of these means is chosen as optimal threshold and the image is converted into binary image based on the obtained threshold value. Binary image thus obtained is eroded with the help of proper structuring element. Watershed segmentation thus carried out on the eroded image gives the effective segmentation results. Different kernels such as corn, rice, and wheat are effectively segmented when the proposed methodology is used. 38 corn kernels, 95 rice kernels, and 32 wheat kernels are used for the experimental purpose. The accuracy of segmentation is as high as 94.7%, 96.8%, and 90.6% for corn, rice, and wheat respectively.

고령화 사회에 접어들면서 거동이 어려운 장애인과 고령자의 개인 교통수단에 대한 수요가 증가하고 있다. 실제로 2017년 기준 전국 전동보장구 보급수는 9만여 대로 지속해서 증가하는 추세다. 하지만 장애인 및 고령자의 판단 능력과 조정 능력은 정상인보다 상대적으로 차이가 있는 관계로 주행 중 사고 발생의 가능성이 크다. 다양한 사고의 원인 중 하나는 도로 노면 상태의 불균형으로 인해 개인 이동 수단 조향 제어의 간섭이다. 본 논문에서는 이 같은 사고를 예방하고자 도로 노면 상태를 고속으로 인지할 수 있는 암호화 형식 의미론적 분할 알고리즘 을 소개한다. 이를 위하여 도로 노면 파손이 포함된 1,500여 장의 학습용 데이터와 150여 장의 테스트용 데이터를 새롭게 구성하였다. 그리고 이를 활용하여 기존의 Encoder와 Decoder 단계 로 구성된 Auto-encoder 방식과 달리 Encoder 단계로 이루어진 심층 신경망을 제안하였다. 이 심층 신경망은 기존의 방식과 비교했을 때 평균 정확도(Mean Accuracy)는 4.45% 증가하였고 파라미터는 59.2% 감소하였으며 연산 속도는 11.9% 향상되었다. 이 같은 고속 알고리즘을 활 용하여 안전한 개인 이동 수단이 확대 적용되길 기대한다.

As we face an aging society, the demand for personal mobility for disabled and aged people is increasing. In fact, as of 2017, the number of electric wheelchair in the country continues to increase to 90,000. However, people with disabilities and seniors are more likely to have accidents while driving, because their judgment and coordination are inferior to normal people. One of the causes of the accident is the interference of personal vehicle steering control due to unbalanced road surface conditions. In this paper, we introduce a encoder type semantic segmentation algorithm that can recognize road conditions at high speed to prevent such accidents. To this end, more than 1,500 training data and 150 test data including road surface damage were newly secured. With the data, we proposed a deep neural network composed of encoder stages, unlike the Auto-encoding type consisting of encoder and decoder stages. Compared to the conventional method, this deep neural network has a 4.45% increase in mean accuracy, a 59.2% decrease in parameters, and an 11.9% increase in computation speed. It is expected that safe personal transportation will be come soon by utilizing such high speed algorithm.

항공 우주 산업의 발전으로 발사체의 연구가 많아지고 있다. 더불어 많은 시험이 진행하고 있고 그에 따라 많은 실패도 발생한다. 발사체는 다른 전자장비보다 위험성이 높기 때문에 실패를 최소화할 수 있어야 한다. 따라서 발사체에 전원을 공급하기 전에 배선을 검증할 필요가 있다. 이에 따라 저항 측정을 통해 배선 상태를 검증하였다. 하지만 이전의 저항측정 방식의 배선 점검은 발사체 회로의 커패시터, 인덕터등의 소자들 때문에 정확한 측정을 할 수 없다. 본 논문에서는 TDR 미터를 이용하여 케이블 및 발사체의 연결 상태에서 임피던스를 측정한다. 측정한 정상상태의 임피던스의 기준 값을 설정하기 위해 Z-Segmentation 알고리즘을 사용한다. Z-Segmentation 알고리즘은 우선 임피던스 파형의 피크 값들을 칼만 필터를 이용하여 찾고 찾은 피크값을 Segmentation을 통해 최종 impedance peak segment를 얻는다. 이런 방법으로 배선의 정상 상태에 대한 기준 값으로 설정하고 이를 기준으로 배선 상태를 판정한다.

development of the aerospace industry is increasing the research of projectiles. In addition, many tests are under way and many failures occur accordingly. Projectiles should be able to minimize failures because they are more dangerous than other electronic equipment. Therefore, it is necessary to verify wiring before powering the projectile. Accordingly, the wiring status was verified by resistance measurements. However, the wiring test of the previous resistance measurement method cannot be accurately measured due to devices such as capacitors and inductors in the projectile circuit. In this paper, impedance is measured in the connection state of cables and projectiles using a TDR meter. The Z-Segmentation algorithm is used to set the reference value for the measured steady state impedance. The Z-Segmentation algorithm first finds the peak values of the impedance waveform using a Kalman filter and obtains the final impedance peak segment through segmentation. In this way, the wiring status is determined based on the reference value for the normal state of the wiring.

의미있는 객체를 배경과 분리하는 영상분할기법은 침입자 경보 시스템, 교통 감시 시스템 등에서 중요한 역할을 담당하며, 일반적으로 공간적 동질성이나 시간적 정보를 이용하는 방법으로 나눌 수 있다. 시간적 정보를 이용하는 방법은 프레임간의 화소값이나 에지성분을 이용한다. 화소값 이용은 간단하며 효과적이나 조명 변화 등이 발생할 경우 움직임 검출이 어렵고 에지성분의 이용은 조명의 영향을 받지 않지만 복잡하며 잡음처리에 어려운 점이 있다. 따라서 본 논문은 카메라가 고정된 감시 시스템에서 화소값 비교와 에지 정보를 이용하여 조명 등의 영향을 최소화하는 움직임 객체 추출 방법을 제안한다. 이는 조명변화와 배경영상의 존재여부에 따라 세 가지 움직임 객체 추출 방법을 달리 적용하며, 투영과 형태 처리 연산자를 사용하는 후처리 과정을 거친 후 움직임 객체를 추출한다. 모의실험 결과 제안알고리즘은 조명변화가 발생하더라도 객체 추출의 결과가 우수함을 보이고 있다.

Surveillance system with the fixed field of view generally has an identical background and is easy to extract and segment a moving object. However, it is difficult to extract the object when the gray level of the background is varied due to illumination condition in the real circumstance. In this paper we propose the segmentation to extract effectively the object in spite of the illumination change. In order to minimize the effect of illumination, the proposed algorithm is composed of three modes to the background generation and the illuminational change. Then the object is finally obtained by using projection and the morphological operator in post-processing. A good seg-mentation performance is demonstrated by the simulation result.

eCRM 시스템은 운영 및 협동 CRM과 분석 CRM으로 나누어진다. 분석 CRM은 데이터 마이닝 기법을 사용하여 데이터를 분석하는 많은 컴포넌트로 구성되어 있다. CRM의 기본적인 기능 중에 하나는 세분화이다. 이는 고객의 획득, 고객의 유지, 고객의 가치 증대의 중요한 비즈니스 프로세스를 지원하는 중추의 역할을 지원한다. 본 논문에서는 eCRM 데이터를 세분화하여 분석하기 위해 홉필드 네트워크 알고리즘을 이용하여 구현하였다. 이 모듈의 개발은 보다 더 효과적이고, 사용자들이 편리하게 사용할 수 있는 환경을 제공한다.

The eCRM(electronic CRM) system which is divided into two parts such as "operational and collaborative CRM" and "analytical CRM". The analytical CRM which is the major part in this paper is consisted of many components which analyzes data using data mining technology. One of the fundamental functions in CRM is customer segmentation, which plays the pivotal role of supporting the key business processes of customer acquisition, customer retention and customer value extension. This paper was implemented eCRM Segmentation Agent basis of Hopfield network algorithm. Through the development of a module, which helps to use existing data environment more effectively, it supports an environment that users can use easily.

본 논문은 평균이동방법과 연결요소방법을 이용하여 도로 영역을 추출하는 알고리즘을 제안하였다. 평균 이동 방법은 중심 모드를 찾기 위한 비모수적 통계 방법으로 컬러 영상을 분할하는데 효율적이다. 일반적으로, 영상의 중·하단에 위치하는 정보를 활용하여 도로의 특징점이 추출된다. 이 특징점과 분할된 컬러 영상을 이용하면, 도로의 영역을 추출할 수 있다. 그러나, 도로의 위치정보와 색상정보만으로 도로영역을 추출할 경우, 잡음과 도로 이외의 영역까지 추출되는 단점이 있다. 본 논문에서는 모폴로지 열기·닫기 연산을 이용하여 잡음을 제거하고, 연결요소 방법을 통하여 가장 큰 영역의 부분만을 추출하여 도로 영역으로 결정하는 방법을 제안한다. 제안된 방법은 실험을 통하여 잡음 제거와 보다 정확한 도로 검출됨을 검증한다.

In this paper, we propose a method for extracting a road area by using the mean-shift method and connected-component method. Mean-shift method is very effective to divide the color image by the method of non-parametric statistics to find the center mode. Generally, the feature points of road are extracted by using the information located in the middle and bottom of the road image. And it is possible to extract a road region by using this feature-point and the partitioned color image. However, if a road region is extracted with only the color information and the position information of a road image, it is possible to detect not only noise but also off-road regions. This paper proposes the method to determine the road region by eliminating the noise with the closing / opening operation of the morphology, and by extracting only the portion of the largest area using a connected-components method. The proposed method is simulated and verified by applying the captured road images.

본 논문에서는 사람 얼굴의 표정을 인식하기 위하여 판단 요소를 선정하고 판단 요소의 변화 상태를 파악 하여 표정을 인식하는 방법을 제안한다. 판단 요소를 선정하기 위하여 이미지 영역 분할 방법을 사용하며, 판단 요 소의 변화율을 이용하여 표정을 판단한다. 표정을 판단하기 위하여 90명의 표정을 데이터베이스화하여 비교하였고, 4 개의 표정(웃음, 화남, 짜증, 슬픔)을 인식하는 방법을 제안한다. 제안한 방법은 시뮬레이션을 실시하여 판단 요소 검 출 성공률과 표정 인식률을 통해 검증한다.

In this paper, we propose a method to recognize the facial expressions by selecting face elements and finding its status. The face elements are selected by using image area segmentation method and the facial expression is decided by using the normal distribution of the change rate of the face elements. In order to recognize the proper facial expression, we have built database of facial expressions of 90 people and propose a method to decide one of the four expressions (happy, anger, stress, and sad). The proposed method has been simulated and verified by face element detection rate and facial expressions recognition rate.

바코드 검사기의 성능에 결정적인 영향을 미치는 것은 입력 영상으로부터 바코드 영역을 추출하는 세그먼테이션 과정이며, 기존의 세그먼테이션 기법에는 여러 가지 문제점이 존재한다. 첫째, 허프 직선 변환 방법은 길이 임계값에 매우 민감하여 임계값을 정하는데 어려움이 있다. 둘째, 모폴로지 변환은 영상을 수축, 팽창하는 과정에서 많은 지연시간이 발생한다. 따라서 본 논문에서는 이러한 바코드 검증에서 지연 현상을 해결하고 주변 영향을 적게 받는 해리스 코너 검출 기법 융합형 바코드 영역 검출 기법을 제안한다. 그리고 본 논문에서 제안한 알고리즘을 검증하기 위해 실제 라인과 유사한 실험 환경을 구성하고, 다양한 크기의 바코드 영상과 다양한 위치에서의 바코드 영역 추출실험을 하였다. 결과적으로 제안 기법은 기존의 알고리즘에 비해 주변 환경이나 임계값 설정의 어려움과 영상 처리의 지연 문제를 해결하였고 모든 테스트 영상에 대해 바코드 영역을 100% 추출하는 성능을 보였다.

A segmentation process extracts an interesting area of barcode in an image and gives a crucial impart on the performance of barcode verifier. Previous segmentation methods occurs some issues as follows. First, it is very hard to determine a threshold of length in Hough Line transform because it is sensitive. Second, Morphology transform delays the process when you conduct dilation and erosion operations during the image extraction. Therefore, we proposes a novel Converged Harris Corner detection-based segmentation method to detect an interesting area of barcode in Data Matrix. In order to evaluate the performance of proposed method, we conduct experiments by a dataset of barcode in accordance with size and location in an image. In result, our method solves the problems of delay and surrounding environments, threshold setting, and extracts the barcode area 100% from test images

컴퓨터 비전 분야의 의미론적 영상 분할(Semantic Image Segmentation) 기술은 이미지를 픽셀 단위로 분할 하여 클래스를 나누는 기술이다. 이 기술도 기계 학습을 이용한 방법으로 성능이 빠르게 향상되는 중이며, 픽셀 단위의 정보를 활용할 수 있는 높은 활용성이 주목받는 기술이다. 그러나 이 기술은 초기부터 최근까지도 계속 '세밀하지 못한 분할'에 대한 문제가 제기되어 왔다. 이 문제는 레이블 맵의 크기를 계속 늘리면서 발생한 문제이기 때문에, 자세한 에지 정보가 있는 원본 영상의 에지 맵을 이용해 레이블 맵을 수정하여 개선할 수 있을 것으로 예상할 수 있었다. 따라서 본 논문은 기존 방법대로 학습 기반의 의미론적 영상 분할을 유지하되, 그 결과인 레이블 맵을 원본 영상의 에지 맵 기반으로 수정하는 후처리 알고리즘을 제안한다. 기존의 방법에 알고리즘의 적용 한 뒤 전후의 정확도를 비교했을 때 평균적으로 약 1.74% 픽셀 정확도와 1.35%의 IoU(Intersection of Union) 정확도가 향상되었으며, 결과를 분석했을 때 성공적으로 본래 목표한 세밀한 분할 기능을 개선했음을 보였다.

Semantic image segmentation technology in the field of computer vision is a technology that classifies an image by dividing it into pixels. This technique is also rapidly improving performance using a machine learning method, and a high possibility of utilizing information in units of pixels is drawing attention. However, this technology has been raised from the early days until recently for 'lack of detailed segmentation' problem. Since this problem was caused by increasing the size of the label map, it was expected that the label map could be improved by using the edge map of the original image with detailed edge information. Therefore, in this paper, we propose a post-processing algorithm that maintains semantic image segmentation based on learning, but modifies the resulting label map based on the edge map of the original image. After applying the algorithm to the existing method, when comparing similar applications before and after, approximately 1.74% pixels and 1.35% IoU (Intersection of Union) were applied, and when analyzing the results, the precise targeting fine segmentation function was improved.