Search

firstly, the skewed image of examination paper is corrected by Hough transform, and then mathematical morphology methods are used to locate the frame lines in table. Then connected component analysis is utilized to locate the form region based on maximum area feature. Finally the form frame line is removed according to the vertical gradient, and the answer image is achieved by analyzing the coordinates of objects. [ABSTRACT FROM PUBLISHER]

Though modern ECG machine with digital-out has been applied for years, paper recordings are still chosen by medical organizations especially in China. But the recording paper is easily broken. These ECG data were necessarily to be extracted and keep the valuable ECG information as digital type for clinical information sharing, online diagnosing and ECG database establishing. A method based on K-means was proposed to extract ECG data from paper recordings. The ECG waveform and the background grid were separated well.105 patients' ECG paper recordings were adopted in the experiment. And the recordings are in different damage level, the paper are in different background color and made by different manufacturers. The result shows that ECG waveform can be extracted precisely and smoothly. The precision rate of RR interval, QRS interval, QT interval, ST slope, and R amplitude from ECG data which are digitalized by the approach in the paper could reach 99%. [ABSTRACT FROM PUBLISHER]

A new method based on an improved canny edge detection algorithm for the runway detection and tracking was presented in this paper. Though the traditional Canny Operator has high edge detection capability, it still can be interfered by grave image noise, as its detection accuracy cannot reach single pixel. This paper presented one method that combined Canny Operator with mean filter to represent the runway edge accurately. Then Hough Transform and Chain-Code was used for runway tracking. Experimental results show that this method can have better detection and tracking effect, besides the processing speed is also improved, that laid a favorable foundation for UAV visual navigation. [ABSTRACT FROM PUBLISHER]

This paper focuses on the research of image segmentation accuracy problem because traditional Sobel operator image segmentation is easy to cause the vagueness of image segmentation, contrast is not apparent, segmentation accuracy is low. Directed against these defects, this paper puts forward an improved Sobel operator 2-d maximum entropy digital image segmentation method. This algorithm firstly carries out image segmentation, according to digital image characteristics, then finds its real edge through the threshold of Sobel edge detection algorithm. Finally applies the threshold value obtained from Sobel edge detection algorithm to 2-d maximum entropy image segmentation. Furthermore, this paper uses different threshold segmentation for digital image goals and objective fringes to solve the problem of segmentation inaccuracy as the result of the local adhesion and stack. Simulation experiments show that the proposed algorithm boasts of its good robustness for image segmentation, high segmentation rate, which has been proven to be an effective and applicable algorithm. [ABSTRACT FROM PUBLISHER]

Choosing one's own viewpoint when watching a video program has long been a desire for viewers. To achieve this goal, view synthesis and depth map generation are two fundamental techniques. View synthesis is a signal processing procedure which creates dense virtual views based on sparse real views. Each object inside a frame is warped to a proper position according to its depth information to form the viewpoint changing perception for viewers. Hence, the correctness of depth map influences the view synthesis quality. To increase the accuracy of depth map, this paper proposes an edge-adaptive block matching scheme cooperated with an unreliable region repairing approach. The former avoid finding local minimum in stereo matching, and the latter repairs the errors caused by occlusion regions. As for view synthesis, this paper proposes a special warping method that can detect errors caused by boundary mismatches of objects between corresponding depth and color images to improve quality of the synthesized view. Besides, we also propose a compensative-filling method that can fix tiny cracks due to round-off errors. Because of these two features, the proposed view synthesis becomes more robust to tolerate errors inside depth maps when compared with previous schemes. Both the depth generation and view synthesis are extremely complex computations. Therefore, this paper also proposes a low-complexity computing technology based on group-of-pixels which increases 30 times of performance for depth map generation, and reduces 60% computation time of view synthesis. [ABSTRACT FROM PUBLISHER]

This paper proved a new adaptive threshold in spherical coordinate system based on Besov space norm theory for application of the internet of things (IOT). It presented a new adaptive curve shrinkage function to overcome the limitation of translational functions. The new function could reach and exceed the true value and enhance the image edge. According to the image statistical characteristics in the spherical coordinates, to process the radial component with the new thresholds and new shrinkage function. The simulation experiments evaluated the new algorithm from the peak signal noise ratio, mean square error and running time. [ABSTRACT FROM PUBLISHER]

In order to make right decision within organizations, individuals and teams need to be supported during decision making (DM) process. The aim in this paper is to support DM by finding out the right people that can be involved to make right decision in more effective and efficient way. In this paper, we explain how it is interesting to use community detection techniques to support DM process. The community can be detected in three phases: the problem formulation and/or the solutions generation and the decision making. To detect community, we propose a simple thresholding algorithm based on Structural/Attribute clustering. [ABSTRACT FROM PUBLISHER]

PET and CT image registration is an important tool of clinical diagnosis of diseases. For PET and CT images, a preprocessing algorithm of medical image registration is proposed in this paper. The algorithm process includes image normalization, CT image adaptive threshold adjustment and automatic extraction of tissues based on morphology, edge detection and statistical analysis theory, and improved PET image interpolation based on physical spacing of pixels. The experiment results indicate that the preprocessing methods proposed in this paper have a great effect on PET and CT image registration research because of the exclusion of redundant information and the improvement of the registration algorithm efficiency. [ABSTRACT FROM PUBLISHER]

By using wavelet transform modulus maxima method to detection image edge, edge details are easily smoothed out in the large scale analysis and related parameters great influenced by the noise is not easy to extract in traditional small scale analysis. To solve this problem, this paper proposes a method based on one-dimensional discrete wavelet image edge detection. This algorithm decompose image into one-dimensional signal, making signal-noise separation with one-dimensional discrete wavelet, and detect the edge of de-noised signal's high frequency components. The article has experimented the multiple vehicle detection in real scene for many times, and the result shows that this algorithm solved the problem that exist in wavelet transform modulus maxima method to test image edges in small scale analysis, restraining noise better, and had higher precision in edge localization. [ABSTRACT FROM PUBLISHER]

This paper presents an image deblurring scheme based on visual saliency. The image is first divided into a significant part and a non-significant part through its saliency map, and the blur kernel of the non-significant part is estimated with an improved fast motion deblurring algorithm. Finally, a compensation mechanism is proposed to convert the space variable blur into the space invariable blur. The predicted blur kernel can be used into the whole image. Experiments show this scheme can effectively deal with the problem of space variable blurring. [ABSTRACT FROM PUBLISHER]

The mean filter can remove the Gaussian noise of an image, and the median filter can remove salt and pepper noise of an image. But two methods both result in some loss of image detail and edge information. In fact, image noise is often mixed, and filters above are not good at removing mixed noise. In this paper an improved adaptive image filter is proposed. According to compare the value of seed pixel and pixels in the neighborhood of seed pixel, the type of noise can be sentenced. Then the median filter can be used to reduce salt and pepper noise. The image edge and detail information can be preserved by the threshold when the mean filter is used to smooth the left pixels. The improved adaptive filter in this paper has combined advantages of the mean filter and the median filter. Experimental results show that the adaptive filter algorithm can effectively filter out both the Gaussian noise and salt and pepper noise of the image, and preserve the image edge and detail information. [ABSTRACT FROM PUBLISHER]

Generally, traditional edge detection methods don't consider the spatial relationship between the adjacent image areas to extract the edges. For the traditional algorithm insufficiency, this paper purposes a novel edge detection algorithm by introducing spatial relationship. This method can be divided into three main steps: firstly, a measure of similarity between pair wise pixels is taken into account by orientation energy. Then, the spatial relationship is needed to find regions where similarity between pixels in a given region is high and similarity between pixels in different regions is low. After that, edge detection is completed with spectral clustering method. Using IKONOS image, the experimental results show that the edge detection method of this paper gained ideal result. [ABSTRACT FROM PUBLISHER]

This paper presents an improved region growing method. Firstly, the automatic seed region growing method is used in this paper to overcome the blindness for selecting the seed pixels. Secondly, the Normalized Cut (N-cut) method is used between the regions to overcome the disadvantage of the region growing without meaningful regions. Finally, the final image segmentation is obtained. The experimental results show that our method can produce good results as favorably compared to some existing algorthms and it is an effective method for image segmentation. [ABSTRACT FROM PUBLISHER]

Traditional image subcategory classification methods combined multiple features into a feature vector. Such methods neglect distinct roles of diverse features on discriminating image subcategories. In this paper, the Dempster-Shafer evidence theory is applied to fuse different features in image subcategory classification. It considers the different contribution of each feature to image classification with limited samples. The experimental results on car subcategory classification show that our proposed method outperforms the k nearest neighbor algorithm in terms of classification accuracy. [ABSTRACT FROM PUBLISHER]

To solve the problem of urban congestion, various vehicles management technology have been developed vigorously, such as the parking management system development is relatively mature, also using of the varied technology. This paper focuses on edge detection which are based on the static image preprocessing. Edge detection makes use of Roberts operator mostly, and the operator has been improved in-depth analysis in this paper. In addition, this paper advanced a new image segmentation algorithm which is based on connected region. [ABSTRACT FROM PUBLISHER]

In this paper, a new registration method that based on shape information and gray information has been introduced to the procedure of CT-MRI registration. In the first, the frames of the two images to be registered were explained by the principle of mechanics decomposition, then we coarsely registered the two images by their frames and as a result the error of the registration was limited in a small region. In the second step, we re-registered the two images that have been coarsely registered by means of maximum mutual information (MMI), and finally an accurate registration result was obtained. The simulation result shows this method is less time consuming and on the other side the registration accuracy is also guaranteed. Finally we successfully applied the present method to register the CT images and MRI images in a patient undergoing neurosurgery. [ABSTRACT FROM PUBLISHER]

Aiming at the identification difficulty of texture-rich veneer defect images, a new variation model based on coupling image texture-structure decomposition and edge extraction is proposed in this paper. Firstly, An advanced AAFC structure-texture decomposition model is obtained through extending the regular items of AAFC model; Secondly, using the semi-quadratic regularization method to obtain a new model of coupled texture extraction and edge detection, and combining with the Chambolle's projection algorithm to finish the numerical solving of the now model, and finally realizing the structure-texture decomposition and extracting the edge information of veneer defect images. The experiment results show that the new model can get better edge information while conducting the structure-texture decomposition of veneer defect images, which indicates that the effect of image edge extraction in this paper can be better than separate edge extraction. [ABSTRACT FROM PUBLISHER]

Conventional edge triggered flip-flops sample a data signal synchronizing with single clock edge. If a noise signal occurs around the clock edge, flip-flops result in malfunction. Then, we have proposed dual edge triggered flip-flops to solve this problem. The flip-flop has highly ability to prevent sampling a noise signal on a data line because it samples the data signal synchronizing with both of the rising edge and the falling edge. In this paper, we design a new circuit of the dual edge triggered flip-flops to improve circuit size, power consumption, and operation speed. In addition, we apply the dual edge triggered flip-flops to signal delay detection. [ABSTRACT FROM PUBLISHER]

Edges Detection is a significant task in many fields of application. Some signal features can be reflected by their discontinuous structure. In general, a mathematical signal transform carries more features than a raw signal. Thus, for analyzing an intercepted signal in a totally blind environment, one can benefit from the information the signal spectrum offers through the Fourier Transform. After sensing the spectrum, the characteristics extraction method must be applied. In this paper, the Wavelet Transform is used, as it exhibits excellent ability to analyze a signal. Also introduced in this paper is an improved algorithm that automatically identifies bandwidth signal boundaries in a determined frequency range. The proposed algorithm is based on the local maxima of the Wavelet Transform modulus. In blind conditions and in a noisy environment, this approach grants good performance. [ABSTRACT FROM PUBLISHER]

Video surveillance systems are becoming very common nowadays. Cameras installed in many places are exposed to sabotage or tampering. This can be done by covering the camera lens, changing the focus of the camera lens or changing the camera position to prevent proper registration of the surveilled area. This paper describes a hardware implementation of a system that can detect these kind of events. The algorithm is based on background modelling, histograms comparison, edges comparison and analysis of the image's average brightness. In was described in a hardware description language in a pipeline manner and implemented in an FPGA device. Real-time processing of a video stream with a resolution of 640×480@60 frames per second was achieved. Tests performed on several sequences demonstrated the usefulness of the presented solution. [ABSTRACT FROM PUBLISHER]

eMTM (electronic Made-to-Measure) is a kind of digital clothing technology that emerges in the 21st century. The core part of it is three-dimensional digital mannequin generation engine. The paper improves the related algorithm of original three-dimensional digital mannequin generating technology by researching mannequin generation method in three-dimensional digital clothing. Particularly, it puts forward the idea that using iterative thresholding method to extract the human body contour for solving shadow problem existing in the photos submitted by the users. Therefore, it can improve the outline continuity of human body image, strengthen the reliability about characteristic data obtained later and enhance the accuracy generating three-dimensional personalized digital mannequin. [ABSTRACT FROM PUBLISHER]

In this paper we present a new architecture for implementing real-time stereo vision on FPGA chips. The proposed architecture is based on reducing the computational needs by focusing on specific image features only instead of processing every image pixel. Two classes of features are considered. The first are low level features like edges and the second are high level features like complete patterns or regions. The paper discusses how both types of features can be integrated with depth calculations to reduce the required FPGA resources while maintaining real-time performance. This allows implementation on relatively small FPGA chips or when limited resources are available. The proposed architecture was successfully implemented on a Virtex 4 FPGA and tested using several sample data sets. The results show that the proposed architecture has excellent accuracy coupled with a significant reduction in required resources. [ABSTRACT FROM PUBLISHER]

After extracting the candidate region from an image, it is necessary to take a kind of technology to determine whether the split target is a pedestrian. By analysis and feature extraction to segmentations of pedestrian candidate region, the classification of pedestrians has been studied. The pedestrian classifier of the SVM (Support Vector Machines) has been trained with pedestrian's typical characteristics. This paper mainly studies the efficient algorithms of splitting pedestrian target from other non-pedestrians. As the pedestrians in the image will show different shapes, postures and sizes, and they are usually in different light conditions, it is complicate to describe the pedestrians. This paper proposes a pedestrian segmentation method, which effectively solves the problems, making the classifier be able to deal with the complicate problems. Secondly, the paper uses the pedestrian image texture and shape features to describe the pedestrian. The extracted features are taken as the input of SVM. In order to solve the impact of lighting and other factors to pedestrian recognition, some characteristics have been considered, such as the pedestrian's grayscale images have certain gray symmetry and texture features, and the pedestrians successive edge makes outline of the image features clear. By using lots of events to train the SVM algorithm the recognized pedestrian classification can be obtained. The test results show that the proposed algorithm can effectively recognize different pedestrians in front of the vehicle and get a good real-time effect. [ABSTRACT FROM PUBLISHER]

In this paper, a circle object recognition method based on monocular vision for the home security robots is proposed. This vision system is able to process image and recognize a colored ball rapidly. The proposed method consists of two submodules, which are the object segmentation module and the circle detection module. In order to implement the object segmentation, the color feature is applied to find out the region of the object. After the region of the object is determined, a fast randomized circle detection method is used by checks if there have enough number radius which are the same in a circle in the region. Because of the double recognition, this system can improve the precision for detecting a colored ball. The proposed method is tested on a home security robot to find a ball. The experimental results illustrate the effectiveness of the proposed method. [ABSTRACT FROM PUBLISHER]

The traditional wavelet edge detection algorithm doesn't ideal processing effects for the salt & pepper noised images. In allusion to this shortcoming, an improved wavelet multi-scale edge detection algorithm is proposed in this paper. Morphology is assimilated into the wavelet detection, and an objective data evaluation is given. Compared with the traditional method, the related coefficient of the improved algorithm is higher, and the distortion degree is reduced. It can effectively filter the salt & pepper noise, improve the accuracy of edge detection, and achieve an ideal effect of edge detection. [ABSTRACT FROM PUBLISHER]

Many methods on extracting 3D information from 2D images have been studied since 1990s, especially the depth information extraction. A novel approach for depth extraction is proposed and implemented in this paper. About 30 images are taken especially for this case study and experiments are conducted on the Stanford Range Image Data. Results show that this approach is generally suitable for most of the images; and performs even better on images with darker background. [ABSTRACT FROM PUBLISHER]

In this paper, a new deinterlacing algorithm which uses an adaptive bilateral filter with local intensity histogram screening and generalized fuzzy operation is proposed. Adaptive bilateral filter with local intensity histogram screening can effectively prevent producing noise magnification or over-sharpening artifacts. Generalized fuzzy operation can enhance image local contrast and prevent involving noise. Experimental results demonstrate the proposed algorithm outperforms the mostly popular algorithms. [ABSTRACT FROM PUBLISHER]

In stock networks analysis, the influential stocks is usually identified by using minimal spanning tree (MST) to filter the important information followed by the centrality measures analysis. In this paper, we introduce an analysis to identify the stocks that might have different behaviour compared to the others. Like the centrality measures analysis, this analysis is also conducted based on MST. A case study on stock networks at KLSE will be reported and discussed. [ABSTRACT FROM PUBLISHER]

Typical image features such as color, shape, and texture are used in content-based image retrieval. Retrieval which uses only one image feature has worse performance in case where the content of an image is complex or where a database contains many images. So, many approaches for integrating these features have been studied. However, the problem of these approaches is how to appropriately assign weighting of the image features at query time. In this paper, we propose a new retrieval method using adaptive image feature weighting. We perform computer simulations in test databases which consist of various kinds of images. The experimental results show that the proposed method has better performance than previous works in respect to several performance evaluations such as precision vs. recall, retrieval efficiency, and ranking measure. [ABSTRACT FROM PUBLISHER]

Community detection is vital to analyze the structure of complex networks. However, existing community detection algorithms didn't consider the function of the community which can always share the common interests and has more interactions amongst its members. In this paper, we presented a novel community detection method by analyzing the spread of behavior in complex networks. We experiment with several real-world networks datasets. The results show that our method is more efficient and flexible which also exhibited great improvement in classifying. [ABSTRACT FROM PUBLISHER]

In this paper, we present the combined character segmentation algorithm based on the active contour model and nonlinear diffusion techniques. The active contour model is used to perform segmentation of printed characters. The coherence enhancing diffusion technique is proposed to smooth out artifacts and background noises without destroying the edges. The performance of the two character segmentation methods: i) the combined ACM-FGM and CED algorithm, and ii) the ACM-FGM algorithm have been validated using a large scale printed documents in Hindi, Malayalam and Telugu text. The combined algorithm achieves an average segmentation accuracy of 89.08% whereas the ACM-FGM algorithm alone had an average accuracy of 52.63%. The whole character segmentation process time is lesser than that of the ACM-FGM algorithm alone. Experiments show that the combined algorithm provides promising results under scanned documents with different font-size and fond-style characters, and the different artifacts and background noises caused by the aging of the paper and diffusion. [ABSTRACT FROM PUBLISHER]

Today, simple and accurate pose estimation is needed for small underwater robots to support and to replace submergence operation divers. Position and orientation estimation using luminous markers and the computer tracking library ARToolKit is one of the solutions for estimation, since it is fast enough for real time processing and is open sourced. Although, since it was made for augmented reality, its use for underwater is limited. In this paper, finding the limit and drawbacks of ARToolKit for underwater usage will be studied. The assessment was done under different levels of turbidities of water and different angles of the marker. The results can be applied to the usage of using ARToolKit underwater with luminous markers. [ABSTRACT FROM PUBLISHER]

The development of embedded system (ES) promotes the trend that the numerical-control equipment becomes more networked and intelligent. ES is the kernal of network and intelligence, yet the image data with contained information is the key point to realize high antomatization and intelligence of ES. The overall automatization level of industrial equipment becomes higher than before. However, in some basic industries, such as the sewing industry, the automatization is backward. This paper analyzes the working mechanism of current pattern-designing software and concludes that the system input is complicated, and the usability and automatization is low. To deal with the problem, the paper proposes a sewing pattern generating method based on optimized image recognition. This method is data-driven, which includes image calibration, edge detecting and curve fitting to simplify the system input. A new algorithm is designed to transfer the system input from vectors to scalars, or to accept both of them, which improves the system's scalability and usability. Corresponding software is developed and the results show that it is more reliable, and it can satisfy the industrial requirements in an acceptable degree. [ABSTRACT FROM PUBLISHER]

Composable Multi-processors employ large instruction windows and distributed layout, both of which amplify the branch misprediction penalty. Once branch misprediction is detected, hundreds or thousands of instructions may be in flight. Simple squashing all the instructions following the mispredicted branch turn to be a large waste. Branch misprediction becomes the key bottleneck in these systems. In this paper, we introduce Distributed Control Independence (DCI) to reduce branch misprediction bottleneck in a composable multi-processor, named TFlex. With control independence, branch misprediction penalty can be alleviated by saving the useful work of future control independent instructions. We found that only a small part of the saving instructions, whose data is depended on control dependent instructions, need re-executing. DCI achieves high hardware efficiency and performance scalability. Our experiment results show that DCI effectively mitigates the bottleneck of branch misprediction and speeds up baseline TFlex by a geometric mean of 35% when running diverse applications on 16-core TFlex configuration. [ABSTRACT FROM PUBLISHER]

This paper proposes a new local self-similarity based backprojection (LSSBP) approach for the image upscaling problem. A backprojection estimates a high resolution image by backprojecting reconstruction error from a low resolution image. The reconstruction error, calculated in the low resolution domain, needs to be anisotropically upscaled in order to reduce artifacts. By using a self-similarity between low and high resolution images, a high frequency data of the high resolution image can be transferred from its similar counterparts in the low resolution domain without many artifacts. However, the direct combination of the backprojection and the self-similarity based upscaling suffers from high computational complexity because similar patches can exist in the entire image and across scales. In this paper, in order to localize search region of similar patch, the local self-similarity based upscaling is adopted for the backprojection of reconstruction error. Therefore, in the backprojection framework, the generated high resolution image has consistency with the low resolution image, and more clean and sharp edge can be obtained by using the local self-similarity. The experimental results show advantages of the proposed approach. The proposed approach can be applied to resolution improvement of surveillance and intelligent vehicle vision systems where images are captured by low resolution sensors. [ABSTRACT FROM PUBLISHER]

Manipulating single cells with a micropipette is the oldest, yet still a widely used technique. This paper discusses the positioning of a single cell to a target position inside the micropipette after the cell is aspirated into the micropipette. Due to the small volume of a single cell (pico-liter) and nonlinear dynamics involved, this task has high skill requirements and is labor intensive in manual operation that is solely based on trial and error and has high failure rates. We present automated techniques in this paper for achieving this task. Computer vision algorithm was developed to track a single cell inside a micropipette for automated single-cell positioning. A closed-loop robust controller integrating the dynamics of cell motion was designed to accurately and efficiently position the cell to a target position inside the micropipette. The system achieved high success rates of 97% for cell tracking (n=100) and demonstrated its capability of accurately positioning a cell inside the micropipette within 8 seconds (vs. 25 seconds by highly skilled operators). [ABSTRACT FROM PUBLISHER]

In order to improve the accuracy and robustness of real-time tracking system, this paper presents new methods for efficient object tracking in video sequences using multiple features and particle filter. Based on the problem that tracking with a single feature is susceptible to interference, the color and edge orientation features are combined under the particle filtering framework, and an adaptive feature-weight assignment approach is also proposed in the process of feature fusion. In the prediction period of particle filter algorithm, the mean-shift method is used to improve the particle swarm optimization algorithm. In this way, the number of effective particles is increased and the real-time performance of the tracking system is improved. Experiment results show that the proposed tracking system is more accurate and more efficient than the traditional color feature based mean-shift algorithm. [ABSTRACT FROM PUBLISHER]

Synthetic Aperture Radar (SAR) systems have been widely used to estimate the various features on the ground. However, the images are often corrupted by noise that can impede further investigation of SAR images. Therefore, the extraction of features from SAR images with noisy backgrounds becomes a challenging issue in SAR image processing. The goal of this paper is to develop and implement a more robust method based on a beamlet transform to extract linear features such as roads from SAR images. The proposed method consists of three steps: First, an image pre-processing technique is used to offset the noise and low-contrast problems by recalculating the pixel values. Second, linear features such as road networks are then extracted by applying a beamlet transform based algorithm. In the third step, a post-processing algorithm is developed to analyze and link the discontinuities in order to connect the road networks in the image. Experimental results have demonstrated the effectiveness of this method. [ABSTRACT FROM PUBLISHER]

At present, micrometer is used for measure the shape edge of automobile airbag. There are some shortcomings in this method. The number of test points is limited. Test data is not comprehensive. Detection speed is slow and a fixture can only test a kind of airbag. The method of airbag shape edge detection based on image processing is researched in this paper. The image of airbag is collected by CCD, and then it is sent to the computer to be processed and segmented. The edge of image is extracted through Canny edge detection algorithm in order to acquire shape edge of airbag in this paper, and the image similarity degree are calculated to provide the information of matching in the template matching process. Finally the comprehensive test shape edge of airbag is realized. The experimental results show that the detection method is effective feasible, intuitive and clear. [ABSTRACT FROM PUBLISHER]

Using coding artificial point can match the homonymy points during multiple images automatically in the photogrammetry. A method of identifying and locating 8 bits encoded points automatically has been presented in this paper. The experimental results show that the accuracy of center coordinate of encoding point reaches sub-pixel level. And decoding method is simple. The robust of decoding and recognition rate is high. So it can meet the requirements of close range photogrammetry fully. [ABSTRACT FROM PUBLISHER]

This paper proposes a multi directions detection deinterlacing algorithm using fixed adaptive tap interpolation filter. The proposed algorithm determines the multi directions of the interpolated pixel, to obtain more direction correlation pixels and use fixed coefficient filter for multi pixels interpolation. The proposed algorithm which adopts only direction correlation pixels can increase pixels correlation and improve visual quality. But multi directions detection costs a little more computation time. Experimental results demonstrate that the proposed algorithm outperforms the conventional deinterlacing methods. [ABSTRACT FROM PUBLISHER]

Depth maps taken from depth cameras are often available at a lower resolution compared to the corresponding color images. In this paper, we propose a selective interpolation method to expand depth image to the size of the corresponding color image. To achieve our goal we selectively adopt either the bilinear smoothing filter or an edge-preserving filter to expand the low resolution depth map. The edge-preserving filters we considered in this paper are joint bilateral up-sampling (JBU) filter and the New Edge-Directed Interpolation (NEDI) filter. Our method not only maintains the visual quality of the interpolated depth image at edge regions, but it also reduces the computational complexity for real time execution. [ABSTRACT FROM PUBLISHER]

This paper introduces a design of clock duty cycle stabilizer (DCS) for high-speed pipelined ADC, and analyses the internal parameters on the impact of the circuit performance. Circuit module includes programmable clock input buffer, clock synthesizer, duty cycle detection circuit and nonoverlapping clock generation circuit. The circuit and layout are achieved by 0.18 μm CMOS 1P5M Mixed Signal process. The Cadence Spectre post-simulation results show: The circuit can work for a wide frequency range from 20MHz to 250MHz; duty cycle accuracy of (50±0.25) %, in the 250MHz input frequency, the RMS jitter is 53 fs. The measured performance shows it can work with high speed, high precision and low jitter characteristics, being not strictly requirement on the input clock signal, nonoverlapping time controllable. [ABSTRACT FROM PUBLISHER]

Edge detection is one of the most used methods for feature extraction in computer vision applications. Feature extraction is traditionally founded on pattern recognition methods exploiting the basic concepts of convolution and Fourier transform. For color image edge detection the traditional methods used for gray-scale images are usually extended and applied to the three color channels separately. This leads to increased computational requirements and long execution times. In this paper we propose a new, enhanced version of an edge detection algorithm that treats color value triples as vectors and exploits the geometric product of vectors defined in the Clifford algebra framework to extend the traditional concepts of convolution and Fourier transform to vector fields. Experimental results presented in the paper show that the proposed algorithm achieves detection performance comparable to the classical edge detection methods allowing at the same time for a significant reduction (about 33%) of computational times. [ABSTRACT FROM PUBLISHER]

This paper presents a Voronoi-based algorithm to extract the medial axis through labeling sample points. A major issue of the medial axis is its inherent instability under small perturbations. The medial axis is very sensitive to small changes of the boundary, which produce many irrelevant branches in the medial axis. Filtering extraneous branches is a common solution to handle this issue, It may be applied as a pre-processing step through simplifying (smoothing) the boundary, or as a post-processing step through pruning, which eliminates the irrelevant branches of the extracted medial axis. However, filtering may alter the topological or geometrical structure of the medial axis. This paper proposes a modification to a Voronoi-based medial axis extraction algorithm to automatically avoid appearing irrelevant branches through labeling the sample points. The experimental results indicate that our method is stable, even in the presence of significant noises and perturbations. [ABSTRACT FROM PUBLISHER]

This paper presents a robust approach for detection of available car parking spaces. With low quality of video camera as webcam and dynamic change of light around the car parking, it is hard to accurately detect or recognize the cars. Moreover the proposed appearance-based approach is efficient than recognition-based approach because it do not need to learn a huge of multi-view objects. In this paper, we propose adaptive background model-based object detection with dynamic mixing features of masked-area and edge orientation histogram (EOH) density. The average variance of variance of intensity change for dynamic background model is used to change ratio of mixing features dynamically. The masked-area density is density of predefined area of a parking slot that is weighted by Gaussian mask to robust density computation and the edge orientation histogram (EOH) density is density of the EOH in the predefined area that can be used under low contrast image as night scene. The experiments are performed both in simulation model and real scenes. The results show the proposed approach can handle dynamic change of light efficiently. [ABSTRACT FROM PUBLISHER]

Owing to the development of 3D industry, the number of 3D contents is growing. However, there does not exist yet a clear, precise criterion to objectively evaluate the quality of these contents. In this paper, we propose a method to measure, among 2D-plus-Depth image quality measure factors, the stereoscopic depth perception. To measure the variety of depth values and the amount of depth-perceived objects, which are assumed to be depth perception-related, we propose two methods, which are based on the histogram and gradient of depth image, respectively. Experimental results show that the proposed methods can accurately estimate depth perception quality. [ABSTRACT FROM PUBLISHER]

Accurate depth map at high resolution is required in many 3D video concepts. Given a low-resolution depth map, this paper studies how to enhance its resolution with a registered high-resolution color image. The idea of the proposed approach is that pixels with similar color values and small distances should have similar depth values, while color discontinuities indicate sharp depth changes at object edges. Therefore, the known depth values in input depth map can be propagated to estimate the unknown depth values of their neighboring pixels with similar color values and small distances in high-resolution depth map. Different from conventional approaches, the proposed approach utilizes the ant colony optimization (ACO) technique to dispatch artificial ants moving on a coupled graph, which consists of a depth map and a color image, and propagate the known depth information from the observed low-resolution depth map to its up-sampled counterpart. Experimental results show that the proposed approach achieves high-resolution depth maps at more desirable quality than that of conventional approaches. [ABSTRACT FROM PUBLISHER]

This paper addresses the problem of Car Make and Model recognition as an example of within-category object class recognition. In this problem, it is assumed that the general category of the object is given and the goal is to recognize the object class within the same category. As compared to general object recognition, this problem is more challenging because the variations among classes within the same category are subtle, mostly dominated by the category overall characteristics, and easily missed due to pose and illumination variations. Therefore, this specific problem may not be effectively addressed using generic object recognition approaches. In this paper, we propose a new approach to address this specific problem by combining global and local information and utilizing discriminative information labeled by a human expert. We validate our approach through experiments on recognizing the make and model of sedan cars from single view images. [ABSTRACT FROM PUBLISHER]

In this paper we propose a novel method of estimating vanishing point by spherical gradient. In contrast with the conventional methods in which vanishing point is estimated from lines, the proposed method does not necessarily extract lines, but employs the spherical gradient cues of edge points. Based on the observation that spherical gradient is aligned with the normal vector of the projection plane of space lines, the vanishing point is estimated directly from spherical gradient of edge points by the Hough Transform. [ABSTRACT FROM PUBLISHER]