Showing papers on "Pixel published in 1990"

Integrating region growing and edge detection

Abstract: A method that combines region growing and edge detection for image segmentation is presented. The authors start with a split-and merge algorithm wherein the parameters have been set up so that an over-segmented image results. Region boundaries are then eliminated or modified on the basis of criteria that integrate contrast with boundary smoothness, variation of the image gradient along the boundary, and a criterion that penalizes for the presence of artifacts reflecting the data structure used during segmentation (quadtree in this case). The algorithms were implemented in the C language on a Sun 3/160 workstation running under the Unix operating system. Simple tool images and aerial photographs were used to test the algorithms. The impression of human observers is that the method is very successful on the tool images and less so on the aerial photograph images. It is thought that the success in the tool images is because the objects shown occupy areas of many pixels, making it is easy to select parameters to separate signal information from noise. >

Unsupervised texture segmentation using Gabor filters

Abstract: A texture segmentation algorithm inspired by the multichannel filtering theory for visual information processing in the early stages of the human visual system is presented. The channels are characterized by a bank of Gabor filters that nearly uniformly covers the spatial-frequency domain. A systematic filter selection scheme based on reconstruction of the input image from the filtered images is proposed. Texture features are obtained by subjecting each (selected) filtered image to a nonlinear transformation and computing a measure of energy in a window around each pixel. An unsupervised square-error clustering algorithm is then used to integrate the feature images and produce a segmentation. A simple procedure to incorporate spatial adjacency information in the clustering process is proposed. Experiments on images with natural textures as well as artificial textures with identical second and third-order statistics are reported. The algorithm appears to perform as predicted by preattentive texture discrimination by a human. >

Super resolution from image sequences

Abstract: An iterative algorithm to increase image resolution is described. Examples are shown for low-resolution gray-level pictures, with an increase of resolution clearly observed after only a few iterations. The same method can also be used for deblurring a single blurred image. The approach is based on the resemblance of the presented problem to the reconstruction of a 2-D object from its 1-D projections in computer-aided tomography. The algorithm performed well for both computer-simulated and real images and is shown, theoretically and practically, to converge quickly. The algorithm can be executed in parallel for faster hardware implementation. >

Solid state color display system and light emitting diode pixels therefor

Abstract: Wide color range display systems comprising integrated, phase modulated light from three light emitting diode (LED) pixels. Each pixel comprises a large number of LED chips arranged compactly to provide a discrete element light source of sufficient output to be viewed as a point source of light from a substantial distance. The arrays of pixels are placed in a matrix of a type typically used in scoreboards, message centers and other large display systems, although the various combinations, subcombinations, and elements are not limited to such uses. Each pixel is mounted in a molded package which may include a transparent lens covering and sufficient number of connecting leads to provide for the number of colors of LEDs contained in the pixel array. Each pixel is placed in a mounting fixture which also accommodates the necessary electrical connections to multiplexed driving circuitry. The light emitted is determined by the type of LED used in the array, preferably an array of red, green and blue, although amber, green and other color combinations may be used. Using three colors, blue, red, green that are controlled by separate driving circuitry provides the capacity to create any color in the spectrum.

Programmable digital video processing system

Abstract: Programmable apparatus for digital processing of video signals from multiple sources converted to digital format to provide real-time multiple simultaneous special video effects and suitable for direct interface to a conventional microcomputer bus such as an Apple Macintosh II NuBus. The apparatus includes a matrix of multipliers to do real-time video processing permitting special effects such as fading between at least two video image sources, as well as a priority resolver to control display on a pixel by pixel basis of more than ten sources based upon dynamically programmable priority. In addition, a programmable multiple range thresholder, a hardware window generator capable of generating multiple simultaneous windows, a color look up table and optional image capture capabilities are provided. The apparatus also provides for a light pen input, genlocking and a range of special video effects including zooming, mosaicing, panning and blending.

A document skew detection method using run-length encoding and the Hough transform

Abstract: As part of the development of a document image analysis system, a method, based on the Hough transform, was devised for the detection of document skew and interline spacing-necessary parameters for the automatic segmentation of text from graphics. Because the Hough transform is computationally expensive, the amount of data within a document image is reduced through the computation of its horizontal and vertical black runlengths. Histograms of these runlengths are used to determine whether the document is in portrait or landscape orientation. A gray scale burst image is created from the black runlengths that are perpendicular to the text lines by placing the length of the run in the run's bottom-most pixel. By creating a burst image from the original document image, the processing time of the Hough transform can be reduced by a factor of as much as 7.4 for documents with gray-scale images. Because only small runlengths are input to the Hough transform and because the accumulator array is incremented by the runlength associated with a pixel rather than by a factor of 1, the negative effects of noise, black margins, and figures are avoided. Consequently, interline spacing can be determined more accurately. >

A multiresolution hierarchical approach to image segmentation based on intensity extrema

Abstract: A computer algorithm which segments gray-scale images into regions of interest (objects) has been developed. These regions can provide the basis for scene analysis (including shape-parameter calculation) or surface-based, shaded-graphics display. The algorithm creates a tree structure for image description by defining a linking relationship between pixels in successively blurred versions of the initial image. The image is described in terms of nested light and dark regions. This algorithm, successfully implemented in one, two, and three dimensions, can theoretically work with any number of dimensions. The interactive postprocessing developed technique selects regions from the descriptive tree for display in several ways: pointing to a branch of the image description tree, specifying by sliders the range of scale and/or intensity of all regions which should be displayed, and pointing (on the original image) to any pixel in the desired region. The algorithm has been applied to approximately 15 computer tomography (CT) images of the abdomen. >

Volume rendering by adaptive refinement

Abstract: Volume rendering is a technique for visualizing sampled scalar functions of three spatial dimensions by computing 2D projections of a colored semi-transparent gel. This paper presents a volume-rendering algorithm, in which image quality is adaptively refined over time. An initial image is generated by casting a small number of rays into the data, less than one ray per pixel, and interpolating between the resulting colors. Subsequent images are generated by alternately casting more rays and interpolating. The usefulness of these rays is maximized by distributing them according to measures of local image complexity. Examples from two applications are given: molecular graphics and medical imaging.

A maximum correlation method for image construction of IRAS survey data

Abstract: An algorithm is presented for the construction of images using linear array data with nonuniform scan coverage of object space and nonuniform detector responses. The algorithm achieves the maximum correlation between adjacent pixels, i.e., the smoothest image, consistent with the data and data uncertainties. For high spatial data density and signal-to-noise ratio, the achievable spatial resolution can exceed the diffraction limit of the optics. The capability of the algorithm is illustrated using 60-micron data from the region centered on the galaxy M101, obtained during the all-sky survey performed by the Infrared Astronomical Satellite. The 60-micron map produced has a resolution of about 36 arcsec and allows the identification of many H II regions by position and aperture photometry for the brighter ones. The achieved resolution is discussed in terms of the a priori estimate of the mean correlation length of the data, the directly measured FWHM in the final image, and the results of aperture photometry of M101 H II regions NGC 5447, 5455, 5461, 5462 and 5471.

Interlace printing process

Abstract: Quality of printed images as produced by an ink-jet printer (10) is improved by staggering applications of ink dots (30) to pixel locations (36) such that overlapping ink dots (30) are printed on successive passes of a printhead (16) and such that swaths (38) are partially printed on overlapping passes of the printhead (16). Quality of multi-colored or multi-shaded images is further improved by grouping pixels (36) into super pixels (42), and applying various combinations of colored ink dots (30) to the various pixels (36) within each super pixel (42) in a staggered sequence.

Embedding the attribute information into a dithered image

Abstract: Attribute information is indispensable data in the collection and storage of a large amount of image information. If these data are stored together with the image, handling of the materials would become simpler and more reliable. This paper proposes a method to embed the attribute information into a dither image by using the difference of the threshold values of the dither matrix when a multilevel image is binarized. This method can embed the attribute information of about 2 kbytes (equivalent to about 25 percent of the data for the image) into a dithered image of 256 × 256 pixels. The embedded information can be displayed or printed separately at any time. This method can restore the dither image almost perfectly by using the property of the ordered dither method.

Method and apparatus for pixel non-uniformity correction

Abstract: An apparatus for reading and correcting an image including: a sensor which receives light and converts it to electrical signals; a digitizing unit to convert the analog image signal to a digital signal; a calibration unit to capture calibration data for each pixel under known illumination conditions and to produce individual correction values for each pixel and a correction unit comprised of a look-up table addressed by both the digital image signal and the correction values which corrects the digital image data.

A fast line finder for vision-guided robot navigation

Abstract: There are two basic ways to improve the speed of a low-level vision algorithm: careful code optimization and selective processing. Reducing the computational effort expended on each pixel reduces the time required to process an image by a constant factor. Selective processing on a limited portion of an image using a focus of attention can decrease overall computation by orders of magnitude. A fast pixel-based algorithm is developed that uses these principles to achieve fast extraction of lines for use in vision-guided mobile robot navigation. It builds upon an algorithm for extracting lines by grouping pixels with similar gradient orientation. It allows parametric control of computational resources required to extract lines with particular characteristics. >

Scanning apparatus using multiple CCD arrays and related method

Abstract: A scanner apparatus using multiple CCD arrays for scanning an image line by line to produce data representative of the image can correct for misalignment of the arrays in the X-axis (scanning) direction and/or Y-axis (feed) direction using a combination of hardware, software, and firmware. To correct for alignment errors in the X-axis direction, the field of view of each of the arrays must overlap. A start pixel is determined for each array based on the degrees of overlap so that a composite scan line of data can be formed from the video image data from each of the arrays. The scan lines are continuously stored in a buffer memory which operates as a ring FIFO buffer. Alignment errors in the Y-axis direction are corrected by software in the host computer which reads separate portions of the scan lines from the buffer memory corresponding to each of the arrays using on read pointers that are set according to the alignment error. The scanner control firmware and software can measure the alignment of the arrays with a simple test target. In addition, the start pixel and read pointers can be changed while displaying the image on a CRT screen so that corrections can be made simultaneously to viewing the image.

Television scan line doubler including temporal median filter

Abstract: A video pixel interpolator generates interpolation pixels from incoming picture image pixels for use within a television picture scan line doubler system. The interpolator includes a temporal median filter for generating an interpolation pixel by selection of the median one of a plurality of temporal pixel samples derived from incoming video, an intra-field averager means for generating an intra-field pixel average, and a switch connected to switch an input to a line doubler of the system between outputs of the temporal median filter and the intra-field averager in response to detection of pixel motion within the picture image. Improved motion detection, and multi-channel processing are also included within the present invention.

Determining watersheds in digital pictures via flooding simulations

Abstract: The watershed transformation is a very powerful image analysis tool provided by mathematical morphology However, most existing watershed algorithms are either too time consuming or insufficiently accurate The purpose of this paper is to introduce a new and flexible implementation of this transformation It is based on a progressive flooding of the picture and it works for n-dimensional images Pixels are first sorted in the increasing order of their gray values Then, the successive gray levels are processed in order to simulate the flooding propagation A distributive sorting technique combined with breadth-first scannings of each gray level allow an extremely fast computation Furthermore, the present algorithm is very general since it deals with any kind of digital grid and its extension to general graphs is straightforward Its interest with respect to image segmentation is illustrated by the extraction of geometrical shapes from a noisy image, the separation of 3-dimensional overlapping particles and by the segmentation of a digital elevation model using watersheds on images and graphs

Fast profilometer for the automatic measurement of 3-D object shapes.

Abstract: A fast technique for automatic 3-D shape measurement is proposed and verified by experiments. The technique, based on the principle of phase measurement of the deformed grating pattern which carries the 3-D information of the measured object, can automatically and accurately obtain the phase map or the height information of a measured object at every pixel point without assigning fringe orders and interpreting data in the regions between the fringe orders. Only one image pattern is sufficient for obtaining the phase map. In contrast to the fast Fourier transform based technique, the technique processes a fringe pattern in the real-signal domain instead of the frequency domain by using demodulation and convolution techniques, can process an arbitrary number of pixel points, and is much faster. Theoretical analysis, simulation results, and experimental results are presented.

A Unified Coding Method of Image and Text Data Using Discrete Orthogonal Transform

Abstract: A unified coding method of image and text data is proposed in this paper. This scheme enables us to embed a text into an image by using some redundant region in the frequency space. Our results show that a text of about 4 k to 8 kbytes can be embedded into the image of 256 × 256 pixels and it retains an S/N ratio of about 30 to 40 dB.

Method and apparatus for coding an image

Abstract: A method for coding an image or other two-dimensional data array to provide a sequence of images having differing spatial frequency content is disclosed The method generates the output images by taking weighted sums of the pixels in the input image The weights are sets of two-dimensional irreducible scaling and wavelet coefficients

3-D data acquisition by Rainbow Range Finder

Abstract: The Rainbow Range Finder (RRF) has the ability to obtain range information for all image pixels with only one frame TV camera imaging during 1/30 s. The authors propose a novel practical calibration scheme which makes range measurement for all visible pixels possible. An experimental system was developed on that basis, and measurement accuracy was evaluated. A curved surface and a plain surface could be measured with sufficient accuracy. The feasibility of a real-time system with an accuracy of 1% in regard to the measurement region size was demonstrated. >

Signal-to-noise efficiency in magnetic resonance imaging

Abstract: The dependence of signal-to-noise ratio (SNR) on the analog filter, the sampling rate and the number and dimensions of voxels is derived for magnetic resonance imaging (MRI). It is shown that the object signal-to-noise ratio scales directly with the voxel volume and the square root of the number of voxels. Defining an efficiency figure of merit as the SNR divided by the square root of the imaging time, it is shown that efficiency is always improved when imaging with the lowest possible resolution (largest voxel dimensions) consistent with viewing the desired anatomical detail. The results directly imply the relative efficiency of 3-D (volume), 2-D (plane), 1-D (line) and 0-D (point) imaging techniques. It is shown that spatial averaging is an inefficient method of noise reduction in MRI. As long as voxel size is maintained constant, one can image as many pixels in the readout direction as desired with no loss in SNR; that is, the number of pixels in the readout direction has no effect on the image SNR. Further, multiple sampling of each phase encoding value (to improve SNR) has no advantage over increasing the number of pixels in the phase encoding direction while leaving the voxel size constant. Some experimental observations are given.

Method for automatically identifying targets in sonar images

Abstract: To detect and classify features in a sonar image comprised of a matrix of pixels each pixel having a known greyness level a set of windows is defined such that each window is comprised of a set of adjacent pixels such that every pixel of the image is included in at least one window. Properties of the greyness level for the pixels in each window are defined to enable selection of those windows having specific features of greyness level in excess of a predetermined threshold. Matched filter correlations are performed on the pixels in the selected windows to identify highlights and shadows. The identifying highlights and shadows are classified as targets, anomalies or background.

Real time optical pre-detection processing of multispectral image data

Abstract: The present invention is an apparatus for forming a spectrally weighted value from received multispectral radiation. The spectral weighting is done optically prior to detection of the received radiation by dispersing the received multispectral radiation into a plurality of wavelength bin areas. This dispersed radiation is then passed through a weighting filter which includes first and second filter elements for each of the wavelength bin areas. The filtered radiation is then converged to corresponding detectors with the spectrally weighted value formed by the difference between the signals of the first and second detectors. A pair of filters, two detectors and a subtracter are employed to produce a generalized weight factor having positive or negative weights. This system can form one or more spectrally weighted values from the radiation from a single pixel, or one spectrally weighted value form the radiation from a plurality of pixels. The addition of a polarization filter permits the formation of spectrally weighted values for various polarizations of a single pixel.

Motion compensation for interlaced digital television signals

Abstract: Motion compensated coding of interlaced digital video signals is provided. Pixel data from successive interlaced video fields is compared to similarly situated pixel data in a plurality of previous fields to find a preferred counterpart for each current field set of pixel data. A prediction error signal is generated for each current field set of pixel data indicative of the difference between it and its preferred counterpart from a previous field. Each prediction error signal is encoded with data indicative of the previous field in which the preferred counterpart can be found. In a preferred embodiment, each set of current field pixel data is compared to a range of similarly situated pixels in each of the previous fields, and the prediction error signal is also encoded with motion vector data indicative of the location of the preferred counterpart in its previous field.

Apparatus and method for compressing and expanding multibit digital pixel data

Abstract: Method and apparatus for converting multibit pixel data to a lesser number of bit pixel data and re-expanding the compressed data. Luminance data for each pixel is established as the 5 most significant bits of the original luminance signal. The chrominance information for groups of pixels is subsampled, and a common chrominance value assigned to each of the pixels in a group. The resulting compressed pixels may be 8 bits wide providing economical possibilities to store the 8 bit wide data. The data is expanded for display by adding lower order data bits to the compressed luminance signal data bits. A subsampled chrominance data signal is appended to the expanded luminance data for display.

A differential method for simultaneous estimation of rotation, change of scale and translation

Abstract: The problem of 2-D motion estimation with application to image sequence coding is considered in this paper. A four-parameter motion model, capable of describing rotation, change of scale and translation simultaneously, is proposed. The four motion parameters are estimated directly from an image sequence without the need for establishing the correspondence between pixels or features. The method is a differential one applied in conjunction with a multi-resolution iteration scheme to enhance its measuring range and efficiency. Experimental results demonstrate that the motion parameters can be estimated accurately by the method with minimal computational effort.

New advancements in charge-coupled device technology - Sub-electron noise and 4096 x 4096 pixel CCDs

Abstract: This paper reports on two new advancements in CCD technology. The first area of development has produced a special purpose CCD designed for ultra low-signal level imaging and spectroscopy applications that require sub-electron read noise floors. A nondestructive output circuit operating near its 1/f noise regime is clocked in a special manner to read a single pixel multiple times. Off-chip electronics average the multiple values, reducing the random noise by the square-root of the number of samples taken. Noise floors below 0.5 electrons rms are reported. The second development involves the design and performance of a high resolution imager of 4096 x 4096 pixels, the largest CCD manufactured in terms of pixel count. The device utilizes a 7.5-micron pixel fabricated with three-level poly-silicon to achieve high yield.

Image registration by recognition of corresponding structures

Abstract: A method for automatic image registration which is characterized by its insensitivity to scaling, rotation, and intensity changes is described. The method is based on similarity assessment of the structures in the images and on a check of their spatial arrangement. Pairs of structures that correspond to each other provide sets of control points to geometric mapping functions. An application of the method to remote-sensing image alignment with a reference map is presented. >