Search

We propose an area-based local stereo matching algorithm for accurate disparity estimation across all image regions. A well-known challenge to local stereo methods is to decide an appropriate support window for the pixel under consideration, adapting the window shape or the pixelwise support weight to the underlying scene structures. Our stereo method tackles this problem with two key contributions. First, for each anchor pixel an upright cross local support skeleton is adaptively constructed, with four varying arm lengths decided on color similarity and connectivity constraints. Second, given the local cross-decision results, we dynamically construct a shape-adaptive full support region on the fly, merging horizontal segments of the crosses in the vertical neighborhood. Approximating image structures accurately, the proposed method is among the best performing local stereo methods according to the benchmark Middlebury stereo evaluation. Additionally, it reduces memory consumption significantly thanks to our compact local cross representation. To accelerate matching cost aggregation performed in an arbitrarily shaped 2-D region, we also propose an orthogonal integral image technique, yielding a speedup factor of 5-15 over the straightforward integration. Index Terms--Cross-based region construction, orthogonal integral images, shape adaptive approximation, stereo matching.

Block matching based motion estimation employs conventionally pixel-based sum of absolute differences (SAD) metric for distortion measure. Many fast matching mechanisms have been developed to lower the computational load for block matching, such as those exploiting block feature based or attribute-based SAD calculations. By examining a subblock-sum-based SAD measure, where the subblock-sums can be considered to be intra-subblock features, we propose both intra- and inter-subblock features based SAD measure for more effective block matching in terms of achieving better coding efficiency. Interestingly, the proposed feature based SAD measure can be interpreted as subblock-based sum of absolute Hadamard transformed differences (SATD). The new features can be constructed in a multilevel structure, which may provide a flexible and scalable means for a tradeoff between computation load and coding efficiency. Encoding results are shown to compare the proposed scheme against other relevant feature based SAD measures. Index Terms--Block matching, feature based matching, integral image, motion estimation, sum of absolute differences (SAD), video coding.

Video deinterlacing can be realized using time-space interpolation filters. To improve video deinterlacing quality with respect to missing pixels on moving diagonal lines, we developed a fuzzy concept that utilizes deinterlacing methods. The proposed algorithm consists of three parts. The first part is fuzzy rule-assisted edge-preserving-based deinterlacing (FED), which has an edge-preserving unit that utilizes fuzzy theory to find the most accurate edge direction and interpolates the missing pixels. Using the introduced gradients in the interpolation, the vertical resolution in the deinterlaced image is subjectively concealed. The second part is weighted fuzzy-reasoning-assisted deinterlacing (WFD), which works in the spatio-temporal domain. In this part, the computed weights are considered and multiplied by the candidate deinterlaced pixels, which successively build approximations of the deinterlaced sequence. The third part of the proposed algorithm is weighted fuzzy switching filtering, which analyzes the suitability of each method on system performance and uses a switching algorithm between FED and WFD. The outcome of image interpolation can be adjusted continuously by varying the setting of the membership function for fuzzy inference. Compared with conventional image interpolation methods, the algorithm presented in this paper provides improved edge quality in the deinterlaced image without the introduction of evident artifacts. Index Terms--Deinterlacing, fuzzy weight, image enhancement, image restoration, motion adaptation.

The paper describes a mixed-mode ASIC composed of a fast readout architecture that interfaces with a matrix of 4096 Monolithic Active Pixel Sensors (MAPS). The matrix has 128 columns and 32 rows of pixels and is divided into 256 regions of 4 x 4 pixels, named macro-pixels (MPs). The chip is an upgrade of a smaller version having 256 pixels that was designed and tested. The two chips were designed via STM 130 nm CMOS technology. The pixel dimension is 50 by 50 [micro][m.sup.2]. The work is aimed at improving the design of MAPS detectors with an on-chip fast sparsification system, for particle tracking, to match the requirements of future high-energy physics experiments. The readout architecture implemented is data driven to extend the flexibility of the system, to be also used in first level triggers on tracks in vertex detectors. Simulations indicate that the readout system can cope with an average hit rate up to 100 MHz/[cm.sup.2] if a master clock of 80 MHz is used, while maintaining an overall efficiency over 99%. Index Terms--Digital readout, MAPS, pixel detector, sparsification.

For dielectric inhomogeneous objects, the perceived reflections are the linear combinations of diffuse and specular reflection components. Specular reflection plays an important role in the fields of image analysis, pattern recognition, and scene synthesis. Several methods for the separation of the diffuse and the specular reflection components have been presented based on image segmentation or local interaction of neighboring pixels. We propose a simple and effective method for specularity removal in a single image on the level of each individual pixel. The chromaticity of diffuse reflection is approximately estimated by employing the concept of modified specular-free image, and the specular component is adjusted according to the criterion of smooth color transition along the boundary of diffuse and specular regions. Experimental results indicate that the proposed method is promising when compared with other state-of-the-art techniques, in both separation accuracy and running speed. OCIS codes: 100.3020, 100.2960, 330.1690.

Problem statement: Image processing applications were drastically increasing over the years. In such a scenario, the fact that, the digital images need huge amounts of disk space seems to be a crippling disadvantage during transmission and storage. So, there arises a need for data compression of images. Approach: This study proposed a novel technique called binary merge coding for lossless compression of images. This method was based on spatial domain of the image and it worked under principle of Inter-pixel redundancy reduction. This technique was taken advantage of repeated values in consecutive pixels positions. For a set of repeated consecutive values only one value was retained. Results: The proposed binary merge coding achieved the compression rate of the brain image was 1.6572479. Comparatively, it is 100% more than the compression rate achieved by standard JPEG. Conclusion/Recommendations: This technique was simple in implementation and required no additional memory area. The experimental results of binary merge coding were compared with standard JPEG and it showed that, the binary merge coding improved compression rate compared to JPEG. The same algorithm can be extending to color images. This algorithm can also used for lossy compression with few modifications. Key words: Huffman coding technique, JPEG, bit plane, data table, INTRODUCTION Data Compression is a technique of encoding information using fewer bits than an un encoded representation would use through specific encoding or compression algorithms (2), (4). All forms of [...]

Problem statement: practical approach of detecting edge map was proposed. Approach: The methodology of this approach was presented and tested in order to select the best value of the operator, used to smooth and get the gradient, and the threshold value used to convert the gray gradient to binary edge map, so a practical value of the threshold and edge operator coefficient was investigated, these values used to calculate the gradient in order to get a better edge-map. Results: While increasing the value of C (constant C the first parameter in our practical approach of detecting the edge), we narrow the range of t and at the same time the value of the suitable t will be increased toward 1. Conclusion: This approach can be used to get the best edge map and to get a clear edge map, which can be used later in image segmentation and object extraction. Keywords: Gradient, detecting, edge detection methodology, original image, edge direction, pixels, INTRODUCTION As we a gray image is a set of pixels which are arranged in a 2D matrix, the intersection of a certain row and column defines the x and [...]

It is possible to introduce tolerancing with the realizability test of Givens and Kostinski [J. Mod. Opt. 40, 471 (1993)] with a minor modification: it suffices to add a special matrix to [G][[M]sup.T][G][M]. Consequently, it is possible to check experimental Mueller matrices with a more flexible tool. Application to the emblematic case of the Mueller images of the free space shows that more than 99% of the pixels are in fact physical, while only 2% initially passed the test. OCIS codes: 110.5405, 110.4280, 120.5410.

Accurate measurement of the position of features in an image is subject to a fundamental compromise: The features must be both small, to limit the effect of nonlinear distortions, and large, to limit the effect of noise and discretization. This constrains both the accuracy and the robustness of image measurements, which play an important role in geometric camera calibration as well as in all subsequent measurements based on that calibration. In this paper, we present a new geometric camera calibration technique that exploits the complete camera model during the localization of control markers, thereby abolishing the marker size compromise. Large markers allow a dense pattern to be used instead of a simple disc, resulting in a significant increase in accuracy and robustness. When highly planar markers are used, geometric camera calibration based on synthetic images leads to true errors of 0.002 pixels, even in the presence of artifacts such as noise, illumination gradients, compression, blurring, and limited dynamic range. The camera parameters are also accurately recovered, even for complex camera models. Index Terms--Camera calibration, imaging geometry, image measurement, high resolution, noise, ray tracing, subpixel.

We present a novel method that provides an accurate and precise estimate of the length of the boundary (perimeter) of an object by taking into account gray levels on the boundary of the digitization of the same object. Assuming a model where pixel intensity is proportional to the coverage of a pixel, we show that the presented method provides error-free measurements of the length of straight boundary segments in the case of nonquantized pixel values. For a more realistic situation, where pixel values are quantized, we derive optimal estimates that minimize the maximal estimation error. We show that the estimate converges toward a correct value as the number of gray levels tends toward infinity. The method is easy to implement; we provide the complete pseudocode. Since the method utilizes only a small neighborhood, it is very easy to parallelize. We evaluate the estimator on a set of concave and convex shapes with known perimeters, digitized at increasing resolution. In addition, we provide an example of applicability of the method on real images, by suggesting appropriate preprocessing steps and presenting results of a comparison of the suggested method with other local approaches. Index Terms--Size and shape, length estimate, perimeter, pixel coverage, gray level, quantization.

Detection of defective pixels in solid-state detectors/sensor arrays has received limited research attention. Few approaches currently exist for detecting the defective pixels using real images captured with cameras equipped with such detectors, and they are ad hoc and limited in their applicability. In this paper, we present a probabilistic novel integrated technique for autonomously detecting the defective pixels in image sensor arrays. It can be applied to images containing rich scene information, captured with any digital camera equipped with a solid-state detector, to detect different kinds of defective pixels in the detector. We apply our technique to the detection of various defective pixels in an experimental camera equipped with a charge coupled device (CCD) array and two out of the four HgCdTe detectors of the UKIRT's wide field camera (WFCAM) used for infrared (IR) astronomy [Astron. Astrophys. 467, 777-784 (2007)]. OCIS codes: 040.1520, 040.3060, 040.1240, 100.5010, 100.2000, 100.2550.

We extend recent results for estimating the parameters of a one-dimensional Gaussian profile to two-dimensional profiles, deriving the exact covariance matrix of the estimated parameters. While the exact form is easy to compute, we provide a set of close approximations that allow the covariance to take on a simple analytic form. This not only provides new insight into the behavior of the estimation parameters, but also lays a foundation for clarifying previously published work. We also show how to calculate the parameter variances for the case of truncated sampling, where the profile lies near the edge of the array detector. Finally, we calculate expressions for the bias in the classical formulation of the problem and provide an approach for its removal. This allows us to show how the bias affects the problem of choosing an optimal pixel size for minimizing parameter variances. OCIS codes: 000.5490, 100.2960, 300.3700.

This paper presents an efficient solution to recovering data pixels of images that have undergone optical and electrical channel impairments in holographic data storage systems. The channel impairments considered include interpixel interference, three types of misalignment, and noise. The proposed misalignment-compensation scheme, consisting of realignment and rate conversion, can effectively eliminate misalignment with more than 84% reduction in additions and 74% reduction in multiplications. In addition, several low-complexity techniques are introduced to reduce the complexity of a two-dimensional maximum a posteriori pixel detection method by up to 95% and do so with negligible degradation in detection performance. OCIS codes: 100.3010, 100.3020, 200.3050, 200.4960, 210.2860, 100.0100.

Computational integral imaging method can digitally provide a series of plane images of three-dimensional (3D) objects. However, the resolution of 3D reconstructed images is dramatically degraded as the distance from the lenslet array increases. In this paper, to overcome this problem, we propose a novel computational integral imaging reconstruction (CIIR) method based on smart pixel mapping (SPM). Since SPM is a computational process in which elemental images recorded at long distances are convertible to ones recorded near lenslet array, this can give us the improved resolution of plane images for 3D objects located at a long distance range from a lenslet array. For the effective use of the SPM-based CIIR method, we design a novel two-stage CIIR method by the combined use of the conventional CIIR and the SPM-based one. The conventional CIIR method is applied over a short distance range, while the SPM-based CIIR is used over a long distance range. We carry out some experiments to verify the performance of the two-stage CIIR system. From the experimental results, the proposed system can provide a substantial gain over a long distance range in terms of the resolution of reconstructed plane images. OCIS codes: 100.6890, 110.2990, 110.4190.

Phase retrieval is a wavefront sensing method that uses a series of intensity images to reconstruct the wavefront. The resolution of phase retrieval testing is limited mainly by the resolution of intensity images captured by CCD cameras. A subpixel phase retrieval method is presented to retrieve the wave field at subpixel resolution by using the information of a sequence of low-resolution images captured along the propagation direction. In this method, the sampling interval lhr the wave field under test is smaller than the CCD pixel size in phase reconstruction. The wave field is recovered at subpixel resolution by utilizing the energy conservation relationship between CCD pixels and their subpixels by the subpixel phase retrieval (SPR) algorithm. Numerical experiments have shown that more than a fourfold resolution enhancement can be achieved. The method has also been studied in some experiments under noisy and off-axis conditions. A mirror surface testing experiment was conducted to demonstrate the performance of SPR in the real world. The results of' these experiments have shown the effectiveness and robustness of this method. SPR allows low-resolution images to be used to retrieve high-resolution wave fields and will be useful in testing wave fields from large objects. OCIS codes: 120.0120, 120.5050.

We investigate the relationship among several popular end-member extraction algorithms, including N-FINDR, the simplex growing algorithm (SGA), vertex component analysis (VCA), automatic target generation process (ATGP), and fully constrained least squares linear unmixing (FCLSLU). We analyze the fundamental equivalence in the searching criteria of the simplex volume maximization and pixel spectral signature similarity employed by these algorithms. We point out that their performance discrepancy comes mainly from the use of a dimensionality reduction process, a parallel or sequential implementation mode, or the imposition of certain constraints. Instructive recommendations in algorithm selection for practical applications are provided. OCIS codes: 100.4145, 100.0100, 100.2960.

Segmentation methods based on pixel classification are powerful but often slow. We introduce two general algorithms, based on sparse classification, for optimizing the computation while still obtaining accurate segmentations. The computational costs of the algorithms are derived, and they are demonstrated on real 3-D magnetic resonance imaging and 2-D radiograph data. We show that each algorithm is optimal for specific tasks, and that both algorithms allow a speedup of one or more orders of magnitude on typical segmentation tasks. Index Terms--Image segmentation, pixel classification, sparse classification, supervised learning.

The reliable detection of the three-dimensional position of arbitrary objects in a scene is a key capability of most animals and one of the most important tasks in machine vision. Today's preferred technical solution is optical time-of-flight (TOF) range imaging, due to its simplicity, its distance resolution, its large and adaptable measurement range, as well as the absence of shadowing problems. In order significantly to extend the application areas of TOF 3-D cameras, in particular for outdoor use, we show how their performance can be improved in all relevant respects: Background light suppression is improved by an order of magnitude by the minimum charge transfer method. Multicamera operation is achieved by a binary pseudo-noise modulation/demodulation technique. This method also avoids all practical ambiguity problems typically encountered with harmonic modulation. Higher temporal demodulation resolution becomes possible with a pixel structure employing lateral electric fields. We have realized such pixels with a commercially available CCD/CMOS process, and our measurement results confirm that gigahertz demodulation imaging is possible. The practicality of all theoretical concepts is demonstrated with a miniaturized TOF 3-D camera platform whose LED array light source is modulated at a typical rate of 20 MHz. Our work contributes, therefore, to opening up new application domains of the soaring optical TOF range imaging techniques. Index Terms--Background light suppression, distance ambiguity, distance resolution, time of flight (TOF), 3-D imaging.

We present an effective method for the pixel-size-maintained reconstruction of images on arbitrarily tilted planes in digital holography. The method is based on the plane wave expansion of the diffraction wave fields and the three-axis rotation of the wave vectors. The images on the tilted planes are reconstructed without loss of the frequency contents of the hologram and have the same pixel sizes. Our method shows good results in the extreme cases of large tilting angles and in the region closer than the paraxial case. The effectiveness of the method is demonstrated by both simulation and experiment. OCIS codes: 090.1995, 100.3010, 100.6890.

A framework for robust foreground detection that works under difficult conditions such as dynamic background and moderately moving camera is presented in this paper. The proposed method includes two main components: coarse scene representation as the union of pixel layers, and foreground detection in video by propagating these layers using a maximum-likelihood assignment. We first cluster into "layers" those pixels that share similar statistics. The entire scene is then modeled as the union of such nonparametric layer-models. An incoming pixel is detected as foreground if it does not adhere to these adaptive models of the background. A principled way of computing thresholds is used to achieve robust detection performance with a prespecified number of false alarms. Correlation between pixels in the spatial vicinity is exploited to deal with camera motion without precise registration or optical flow. The proposed technique adapts to changes in the scene, and allows to automatically convert persistent foreground objects to background and reconvert them to foreground when they become interesting. This simple framework addresses the important problem of robust foreground and unusual region detection, at about 10 frames per second on a standard laptop computer. The presentation of the proposed approach is complemented by results on challenging real data and comparisons with other standard techniques. Index Terms--Video analysis, scene analysis, foreground detection, surveillance, background subtraction, layer tracking.

Conventional imaging techniques adopt a rectilinear sampling approach, where a finite number of pixels are spread evenly across an entire field of view (FOV). Consequently, their imaging capabilities are limited by an inherent trade-off between the FOV and the resolving power. In contrast, a foveation technique allocates the limited resources (e.g., a finite number of pixels or transmission bandwidth) as a function of foveal eccentricities, which can significantly simplify the optical and electronic designs and reduce the data throughput, while the observer's ability to see fine details is maintained over the whole FOV. We explore an approach to a foveated imaging system design. Our approach approximates the spatially variant properties (i.e., resolution, contrast, and color sensitivities) of the human visual system with multiple low-cost off-the-shelf imaging sensors and maximizes the information throughput and bandwidth savings of the foveated system. We further validate our approach with the design of a compact dual-sensor foveated imaging system. A proof-of-concept bench prototype and experimental results are demonstrated. OCIS codes: 110.0110, 220.4830, 330.7338, 330.1800, 230.4685.

Missing data are common in range images, due to geometric occlusions, limitations in the sensor field of view, poor reflectivity, depth discontinuities, and cast shadows. Using registration to align these data often fails, because points without valid correspondences can be incorrectly matched. This paper presents a maximum-likelihood method for registration of scenes with unmatched or missing data. Using ray casting, correspondences are formed between valid and missing points in each view. These correspondences are used to classify points by their visibility properties, including occlusions, field of view, and shadow regions. The likelihood of each point match is then determined using statistical properties of the sensor, such as noise and outlier distributions. Experiments demonstrate high rates of convergence on complex scenes with varying degrees of overlap. Index Terms--Range data, registration, maximum-likelihood, pixel classification.

In this paper, we propose a motion-compensated spatio--temporal Locally adaptive linear minimum mean squared-error (LLMMSE) filter for noisy image sequences with both temporal and spatially adaptive filtering support. Motion compensation and an adaptive temporal filtering support (TFS) structure guarantee the uniformity of the TFS. An intelligent pixel aggregation algorithm is proposed to include homogeneous neighboring pixels and exclude the outlier pixels, resulting in uniform spatial filtering support. By using the proposed spatio--temporal LLMMSE filter with uniform spatio--temporal support, we can reduce noise efficiently without introducing visually disturbing blurring artifacts. Furthermore, we employ an adaptive weighted local mean and variance estimation algorithm to improve the accuracy of estimation. The weights provide an implicit mechanism for deemphasizing the contribution of the outlier pixels which are wrongly kept within the support to avoid blurring. The performance of the proposed filter is quantitatively evaluated and compared with state-of-the-art methods. The results demonstrate that the proposed filter can achieve superior filtering performance, both subjectively and objectively. Index Terms--Locally adaptive linear minimum mean squared-error (LLMMSE), motion-compensated spatio--temporal filter, spatial filtering support (SFS), temporal filtering support (TFS).

Subpixel detection is a challenging problem in hyperspectral imagery analysis. Since the target size is smaller than the size of a pixel, detection algorithms must rely solely on spectral information. A number of different algorithms have been developed over the years to accomplish this task, but most detectors have taken either a purely statistical or a physics-based approach to the problem. We present two new hybrid detectors that take advantage of these approaches by modeling the background using both physics and statistics. Results demonstrate improved performance over the well-known AMSD and ACE subpixel algorithms in experiments that include multiple targets, images, and area types--especially when dealing with weak targets in complex backgrounds. Index Terms--Target detection, subspace detectors, hyperspectral data, spectral mixture models.

A prototype pixel detector readout chip has been developed with a new front-end architecture aimed at eliminating the spectral distortion produced by charge diffusion in highly segmented semiconductor detectors. In the new architecture neighbouring pixels communicate with one another. At the corner of each pixel summing circuits add the total charge deposited in each sub-group of 4 pixels. Arbitration logic assigns a hit to the summing circuit with the highest charge. In the case where incoming X-ray photons produce fluorescence--a particular issue in high-Z materials--the charge deposited by those fluorescent photons will be included in the charge sum provided that the deposition takes place within the volume of the pixels neighbouring the initial impact point. The chip is configurable such that either the dimensions of each detector pixel match those of one readout pixel or detector pixels are 4 times greater in area than the readout pixels. In the latter case event-by-event summing is still possible between the larger pixels. As well as this innovative analog front-end circuit, each pixel contains comparators, logic circuits and two 15-bit counters. When the larger detector pixels are used these counters can be configured to permit multiple thresholds in a pixel providing spectroscopic information. The prototype chip has been designed and manufactured in an 8-metal 0.13 [micro]m CMOS technology. First measurements show an electronic pixel noise of ~ 72 e-rms (Single Pixel Mode) and ~ 140 e-rms (Charge Summing Mode). Index Terms--Charge sharing, CMOS, Medipix, photon counting, pixel, X-rays.

A procedure for deconvolution of multiple images of the same object with space-variant point-spread functions (PSFs) is presented. It is based on expressing deconvolution with inverse filtering as convolution with kernels corresponding to inverse PSFs. Sets of basis functions are made from these inverse PSFs, given at discrete sample points, through Karhunen--Loeve (K-L) decomposition. The entire field of view can then be convolved with the K-L kernels. Coadding the results using continuous maps of expansion weights, interpolated for every pixel between the sample points, results in an image that is deconvolved with smoothly varying PSFs that match the discrete measurements. A demonstration data set is used to show how the transition between the grid points improves deconvolutions compared to piecewise deconvolution and mosaicking by avoiding the blending of discontinuities at the interfaces between adjacent subfields.

A displayed image is NOT a document. A displayed image has imaginary inches and is an imaginary document. It does NOT need settings as a document. It needs ONLY settings [...]