• Journal of the Optical Society of Korea
• /
• v.14 no.4
• /
• pp.376-382
• /
• 2010

We propose a novel method to precisely detect car license plates by locating main characters, which are printed with large font size. The regions of the main characters are directly detected without detecting the plate region boundaries, so that license regions can be detected more precisely than by other existing methods. To generate a binary image, multiple thresholds are applied, and segmented regions are selected from multiple binarized images by a criterion of size and compactness. We do not employ any character matching methods, so that many candidates for main character groups are detected; thus, we use a neural network to reject non-main character groups from the candidates. The relation of the character regions and the intensity statistics are used as the input to the neural network for classification. The detection performance has been investigated on real images captured under various illumination conditions for 1000 vehicles. 980 plates were correctly detected, and almost all non-detected plates were so stained that their characters could not be isolated for character recognition. In addition, the processing time is fast enough for a commercial automatic license plate recognition system. Therefore, the proposed method can be used for recognition systems with high performance and fast processing.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.6
• /
• pp.2302-2316
• /
• 2015

We propose an efficient method of extracting targets within a region of interest in non-homogeneous infrared images by using a principal component analysis (PCA) plane and adaptive Gaussian kernel. Existing approaches for extracting targets have been limited to using only the intensity values of the pixels in a target region. However, it is difficult to extract the target regions effectively because the intensity values of the target region are mixed with the background intensity values. To overcome this problem, we propose a novel PCA based approach consisting of three steps. In the first step, we apply a PCA technique minimizing the total least-square errors of an IR image. In the second step, we generate a binary image that consists of pixels with higher values than the plane, and then calculate the second derivative of the sum of the square errors (SDSSE). In the final step, an iteration is performed until the convergence criteria is met, including the SDSSE, angle and labeling value. Therefore, a Gaussian kernel is weighted in addition to the PCA plane with the non-removed data from the previous step. Experimental results show that the proposed method achieves better segmentation performance than the existing method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.12
• /
• pp.5073-5086
• /
• 2015

The rock fracture detection by image analysis is significant for fracture measurement and assessment engineering. The paper proposes a novel image segmentation algorithm for the centerline tracing of a rock fracture based on Hessian Matrix at Multi-scales and Steger algorithm. A traditional fracture detection method, which does edge detection first, then makes image binarization, and finally performs noise removal and fracture gap linking, is difficult for images of rough rock surfaces. To overcome the problem, the new algorithm extracts the centerlines directly from a gray level image. It includes three steps: (1) Hessian Matrix and Frangi filter are adopted to enhance the curvilinear structures, then after image binarization, the spurious-fractures and noise are removed by synthesizing the area, circularity and rectangularity; (2) On the binary image, Steger algorithm is used to detect fracture centerline points, then the centerline points or segments are linked according to the gap distance and the angle differences; and (3) Based on the above centerline detection roughly, the centerline points are searched in the original image in a local window along the direction perpendicular to the normal of the centerline, then these points are linked. A number of rock fracture images have been tested, and the testing results show that compared to other traditional algorithms, the proposed algorithm can extract rock fracture centerlines accurately.

• Journal of Sensor Science and Technology
• /
• v.27 no.4
• /
• pp.216-220
• /
• 2018

In this paper, we present visual sensing techniques that can be used to teach a robot using a laser pointer. The light spot of an off-the-shelf laser pointer is detected and its movement is tracked on consecutive images of a camera. The three-dimensional position of the spot is calculated using stereo cameras. The light spot on the image is detected based on its color, brightness, and shape. The detection results in a binary image, and morphological processing steps are performed on the image to refine the detection. The movement of the laser spot is measured using two methods. The first is a simple method of specifying the region of interest (ROI) centered at the current location of the light spot and finding the spot within the ROI on the next image. It is assumed that the movement of the spot is not large on two consecutive images. The second method is using a Kalman filter, which has been widely employed in trajectory estimation problems. In our simulation study of various cases, Kalman filtering shows better results mostly. However, there is a problem of fitting the system model of the filter to the pattern of the spot movement.

• Korean Journal of Remote Sensing
• /
• v.36 no.5_2
• /
• pp.929-937
• /
• 2020

In this manuscript, the UNet++ model, which is one of the representative deep learning techniques for semantic segmentation, was used to detect changes in temporal satellite images. To analyze the learning results according to various loss functions, we evaluated the change detection results using trained UNet++ models by binary cross entropy and the Jaccard coefficient. In addition, the learning results of the deep learning model were analyzed compared to existing pixel-based change detection algorithms by using WorldView-3 images. In the experiment, it was confirmed that the performance of the deep learning model could be determined depending on the characteristics of the loss function, but it showed better results compared to the existing techniques.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.10 no.5
• /
• pp.871-876
• /
• 2006

In this thesis, to accurately measure 3D face information using the spatial encoding patterns, the new algorithm to segment the pattern images from initial face pattern image is proposed. If the obtained images is non-homogeneous texture and ambiguous boundary pattern, the pattern segmentation is very difficult. Furthermore. the non-encoded areas by accumulated error are occurred. In this thesis, the FCM(fuzzy c-means) clustering method is proposed to enhance the robust encoding and segmentation rate under non-homogeneous texture and ambiguous boundary pattern. The initial parameters for experiment such as clustering class number, maximum repetition number, and error tolerance are set with 2, 100, 0.0001 respectively. The proposed pattern segmentation method increased 8-20% segmentation rate with conventional binary segmentation methods.

• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 1996.06c
• /
• pp.773-779
• /
• 1996

Dried oak mushroom have complex and various visual features. Grading and sorting of dried oak mushrooms has been done by the human expert. Though actions involved in human grading looked simple, a decision making underneath the simple action comes from the result of the complex neural processing of the visual image. Through processing details involved in human visual recognition has not been fully investigated yet, it might say human can recognize objects via one of three ways such as extracting specific features or just image itself without extracting those features or in a combined manner. In most cases, extracting some special quantitative features from the camera image requires complex algorithms and processing of the gray level image requires the heavy computing load. This fact can be worse especially in dealing with nonuniform, irregular and fuzzy shaped agricultural products, resulting in poor performance because of the sensitiveness to the crisp criteria or specific ules set up by algorithms. Also restriction of the real time processing often forces to use binary segmentation but in that case some important information of the object can be lost. In this paper, the neuro net based real time recognition algorithm was proposed without extracting any visual feature but using only the directly captured raw gray images. Specially formated adaptable size of grids was proposed for the network input. The compensation of illumination was also done to accomodate the variable lighting environment. The proposed grading scheme showed very successful results.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 1998.06b
• /
• pp.111-115
• /
• 1998

The Internet today is teemed with not only text data but also other media such as sound, still and moving images in a variety of formats. Unlike text, however, that can be retrieved easily with the help of numerous search engines, there has been few way to access data of other media unless the exact location or the URL is known. Multimedia data in the WWW are contained in or linked via anchors in the hyper-documents. They can most reliably be retrieved by analyzing the binary data content, which is far from being practical yet by the current state of the art. Instead we present another technique of searching based on textual descriptions which are found at or around the multimedia objects. The textual description used in this research includes file name (URL), anchor text and its context, alternative descriptions found in ALT HTML tage. These are actually the clues assumedly relevant to the contents. Although not without a possibility of missing or misinterpreting images and sounds, the description-based search is highly practical in terms of computation. The prototype search engine will soon be deployed to the public service through the prestige search engine, InfoDetective, in Korea.

• 한국정보디스플레이학회:학술대회논문집
• /
• 2005.07a
• /
• pp.575-579
• /
• 2005

A beautiful, flexible active-matrix electrophoretic display (AM-EPD) device is reported. The flexible AM-EPD device has a $40.0{\times}30.0\;mm^2$ display area, measures about 0.27 mm in thickness, weighs about 0.45 g and possesses only 20 external connections. The flexible AM-EPD device displays clear black-and-white images with 5 gray-scales on $160{\times}120$ pixels. The display is free from residual image problems, because we use an area-gray-scale method on $320{\times}240$ EPD elements, each of which is driven with binary signals. Each pixel consists of 4 EPD elements. In addition, since the response time of the electrophoretic material is as long as approximately 400 ms and since the display possesses a large number of EPD elements, we have developed a special driving method suitable for changing EPD images comfortably. A complete image is formed on the AM-EPD device, consisting of a reset frame and several, typically 6, image frames.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.10 no.2
• /
• pp.69-75
• /
• 1985

The proposed seal identification procedure consists of the smoothing, rotation, thinning, and matching techniques. The seal images which are scanned by CCTV are thresholded into the binary prctures of $256{\times}256$ pixels through A/D converter and 6502 microcomputer. After the sample and target images are ratated into an identical orientation, a thinning process is used to extract the skeletons of the character strobes. The wighted map is constructed by distance weight from which the distance weighted correlation C is computed. The C is compared with the dicision constant C or C for the purpose of seal indentification. The identification rate is 95% and the total CPU time is less than 3 minutes for each identification in the experiment.