Showing papers on "Bilateral filter published in 1995"

A multidimensional nonlinear edge-preserving filter for magnetic resonance image restoration

Abstract: The paper presents a multidimensional nonlinear edge-preserving filter for restoration and enhancement of magnetic resonance images (MRI). The filter uses both interframe (parametric or temporal) and intraframe (spatial) information to filter the additive noise from an MRI scene sequence. It combines the approximate maximum likelihood (equivalently, least squares) estimate of the interframe pixels, using MRI signal models, with a trimmed spatial smoothing algorithm, using a Euclidean distance discriminator to preserve partial volume and edge information. (Partial volume information is generated from voxels containing a mixture of different tissues.) Since the filter's structure is parallel, its implementation on a parallel processing computer is straightforward. Details of the filter implementation for a sequence of four multiple spin-echo images is explained, and the effects of filter parameters (neighborhood size and threshold value) on the computation time and performance of the filter is discussed. The filter is applied to MRI simulation and brain studies, serving as a preprocessing procedure for the eigenimage filter. (The eigenimage filter generates a composite image in which a feature of interest is segmented from the surrounding interfering features.) It outperforms conventional pre and post-processing filters, including spatial smoothing, low-pass filtering with a Gaussian kernel, median filtering, and combined vector median with average filtering. >

Apparatus and method for 2-dimensional data compression

Abstract: A system and method for compressing data, including image data. The image data is formed by a pixel array, the pixel array having target pixels and prior pixels, each prior pixel being located in a position within the pixel array prior to each of the target pixels. The method of the present invention includes compressing the image data to obtain a compressed image. In the compression step, the pixel array is traversed according to a predetermined non-linear two-dimensional traversing pattern. A longest matching prior pixel string, if any, is located for each string of target pixels having a string of matching prior pixels. If no matching prior pixel is found for a target pixel, each such target pixel is characterized as an unmatched pixel. Finally, the compressed image is encoded and may be either transmitted or stored.

Structure and method for reducing spatial noise

Abstract: A digital filter for noise reduction selects between local variances obtained from adjacent pixels in the same frame and adjacent pixels in the same field. In one embodiment, the digital filter includes a filter modified from an adaptive Wiener filter which preserves edges and smoothes smooth areas of the image. A high compression ratio can be achieved in very smooth regions of the image without introducing artifacts.

Method and apparatus for compressing and decompressing a video image

Abstract: Edge or contour information is extracted from an image array by filtering and encoded. In order to improve reproduction accuracy, two separate filters (302, 304) are used to detect edge information. One of the filters detects 'sharp' edges, or edges in which the discontinuity in pixel intensity values occurs over a range of a few pixels. The other filter detects 'level' edges in which the pixel intensity value discontinuity occurs over a larger range of pixels than the 'sharp' edges. The 'smooth' areas between edges or contours are assumed to vary continuously between the contours, but for efficient implementation, a one-dimensional linear interpolation (216) is used to regenerate the contour information between edges. In addition, to further improve performance, a line is fitted to the pixel intensity data. The end values of this line are then used for the pixel intensity data. In order to still further improve performance, the pixel intensity values associated with each contour are divided into groups and each group is then encoded. Another improvement is accomplished by mean coding the residual error.

A Ring Median Filter for Digital Images

Abstract: The ring filter is defined as a median filter which assigns weight only to select pixels in an annulus. Its advantage is that it has a sharply-defined scale length; that is, all objects with a scale-size less than the radius of the ring are filtered and replaced by the local background level. It provides a fast, simple and intuitive method to remove the small-scale objects (independent of morphology) from a digital image, leaving behind the large-scale objects and overall light gradients. The ring filter is much faster than the more commonly used filled-box median filter, and completes in one or two passes what previously required a long iterative procedure. Several examples of its use are presented.

Dilation of images without resolution conversion to compensate for printer characteristics

Abstract: An image compensation system which provides dilation or erosion of image features using halfbitting or fullbitting in the rendition of bitmap images, especially on a write-white printer. A region of pixels of an image is isolated which includes two or more correctable pixel locations. A set of state determination rules, based on the formation of pixels in the isolated region, is used to determine a corrected binary pixel state for each of the correctable pixels. Corrections for one correctable pixel may be considered in the state determination rules for adjacent correctable pixels. A single enhanced output pixel is provided for each image input pixel, thereby preserving the original image resolution. Performing enhancements on multiple input pixels simultaneously increases the system throughput.

Enhancement of JPEG coded images by adaptive spatial filtering

Abstract: The JPEG coder has proven to be extremely useful in coding image data. For low bit-rate image coding (0.75 bit or less per pixel), however, the block effect becomes very annoying. The edges also display 'wave-like' appearance. An enhancement algorithm is proposed to enhance the subjective quality of the reconstructed images. First, the pixels of the coded image are classified into three broad categories: (a) pixels belonging to quasi-constant regions where the pixel intensity values vary slowly, (b) pixels belonging to dominant-edge (DE) regions which are characterized by few sharp and dominant edges and (c) pixels belonging to textured regions which are characterized by many small edges and thin-line signals. An adaptive mixture of some well-known spatial filters which uses the pixel labeling information for its adaptation is used as the adaptive optimal spatial filter for image enhancement. Some experimental results are also provided to demonstrate the success of the proposed scheme.

Pixel-independent ray tracing

Abstract: Publisher Summary This chapter discusses the pixel-independent ray tracing. A standard practical reconstruction filter is defined. The rationale for choosing this curve is that the ideal band pass filter is infinite in width, and oscillates. It gives ideal reconstruction from samples taken from a perfectly prefiltered image, but one need a good reconstruction from ray samples. Oscillations produce haloes around sharp edges. Having chosen a standard filter, the next stage is to imagine that one is going to create a continuous image from a set of ray samples, by convolution with that filter. It is this image that one will represent by a grid of pixels. Each ray sample contributes to the continuous image, a fuzzy spot whose intensity is the given cubic. It is found that if the pixel value represents the integral of this image over a square, then one must integrate the cubic over the pixel area to find the contribution of each pixel. One could create images directly using this technique but calculating the filter integrals on the fly is very expensive, so one wrote a program to create tables of the integrals.

System and method for enhancement of coded images using adaptive spatial filtering

Abstract: Pixels in a low bit-rate encoded image are categorized into two or more categories. In one embodiment three pixel categories are used: 1) pixels belonging to quasi-constant (QC) regions where the pixel intensity values vary slowly (e.g., pictorial portions), 2) pixels belonging to textured regions which are characterized by many small edges and thin-line signals (e.g., text portions), and 3) pixels belonging to dominant-edge (DE) regions which are characterized by few sharp and dominant edges (e.g., edge portions). In one embodiment DE pixels are categorized first, then QC and textural pixels are distinguished from the remaining pixels using the number of zero-crossings among the pixels and a threshold. Conventional spatial filters that are well suited for each pixel category type are then used to enhance each region of the image. In one embodiment, various combinations of spatial filters are used to enhance the image.

Non-uniform clocking system for area imaging device

Abstract: An area imaging device comprises an area image sensor and means to adjust the vertical and horizontal clocking rate used to transfer pixels from an array of photo-sensors to a serial output of the area image sensor. The non-uniform clocking rates of the present invention reduce the amount of memory necessary to implement pan and zoom features by effectively separating image and non-image pixels during scanning. This also allows images collected in one format to be displayed on a display device designed for a second format. For each image scan, the area image sensor is divided into active, recovery, and inactive regions according to whether the pixels within the region are displayed, adjacent to displayed pixels, or neither adjacent to image pixels nor displayed, respectively. Rows of pixels are transferred to a horizontal shift register at a vertical scanning rate which is increased for rows that include no image pixels. The pixels of each row are then clocked out of the horizontal shift register at a rate that is adjusted according to the region of the area image sensor in which the pixels originated. Non-image pixels are shifted out at a high rate, while image pixels are shifted out at a rate that minimizes distortion of the image. Recovery pixels which segregate image and non-image pixels, are shifted out of the horizontal shift register at a slow rate to flush charge transfer cells of any accumulated excess charges.

Novel filter algorithm for removing impulse noise in digital images

Abstract: In this paper, we present a novel filter algorithm that is more capable in removing impulse noise than some of the common noise removal filters The philosophy of the new algorithm is based on a pixel identification concept Rather than processing every pixel in a digital image, this new algorithm intelligently interrogates a subimage region to determine which are the 'corrupted' pixels within the subimage With this knowledge, only the 'corrupted' pixels are eventually filtered, whereas the 'uncorrupted' pixels are untouched Extensive testing of the algorithm over a hundred noisy images shows that the new algorithm exhibits three major characteristics First, its ability in removing impulse noise is better visually and has the smallest mean-square error compared with the median filter, averaging filter and sigma filter Second, the effect of smoothing is minimal As a result, edge and line sharpness is retained Third, the new algorithm is consistently faster than the median filter in all our test cases In its current form, the new filter algorithm performs well with impulse noise© (1995) COPYRIGHT SPIE--The International Society for Optical Engineering Downloading of the abstract is permitted for personal use only

Method of extracting pixels constituting intensity changes in an image and method of identifying collection of intensity changes

Abstract: A method for extracting pixels constituting an intensity change formed of a collection of local intensity changes present in a gray image of an object based on a intensity data of each pixel includes the steps of: photographing the object by an image photographing device to obtain the gray image of the object represented by the intensity data of pixels; setting a partial image with a predetermined shape and a predetermined size in the image which has a target pixel whose entire periphery is surrounded by adjacent pixels; calculating intensity data belonging to a specific area composed of the target pixel and peripheral pixels and intensity data belonging to its remaining area for the partial image; obtaining a ratio of the calculating results between the intensity data belonging to the specific area and the intensity data of the remaining area, and comparing the obtained ratio with a threshold value, and then determining suitability of the target pixel as a composing factor of the intensity change based on a compared result, and then extracting pixels constituting the intensity change which have been determined to be suitable.

Coding apparatus for bi-level images halftoned with error diffusion algorithm

Abstract: It is sought to provide a coding apparatus which permits high efficiency coding of bi-level images with black-and-white pixels halftoned with error diffusion algorithm. An image input/pixel selection unit selects, from an input image as a bi-level image with black-and-white pixels halftoned with error diffusion algorithm, first selected pixels with respect to a pixel "?" to be coded and second selected pixels as peripheral pixels with respect to the first selected pixels. The first selected pixels are fed to a pixel pattern binary scale translator for translation into a binary number. The second selected pixels, on the other hand, is fed to a pixel value adder for counting of black pixels. The count is fed to a binary scale translator for translation into a binary number. The binary number outputs of the pixel pattern binary scale translator and the binary scale translator are combined to be input to an address input terminal of a ROM having a coding information table.

Image smoothing based on pixel grouping

Abstract: In this paper we describe an image smoothing method that is capable of reducing both Gaussian noise and impulse noise. For each pixel in an image we examine a window area that is centered at the pixel. Pixels within the window are then divided into three groups based on their intensities relative to the center pixel. One group has higher intensities. One group has lower, and the third group has intensities close to that of the center pixel. A new intensity value for the center pixel is then determined by averaging the pixels in one or more groups based on the population of the three groups. Therefore, the number of pixels used in averaging and the configuration of these pixels may vary from point to point. This high adaptability enables the approach to preserve the sharpness of edges, corners, lines, and other image features in the smoothing process.

Two-tone image data processor e.g. for facsimile

Abstract: The device processes two-tone image data in order to enlarge a relevant image (picture) and to smooth a boundary line between one zone consisting of pixels of a first tone of the two-tone pixels which represent the relevant image, and a zone consisting of pixels of a second tone of the two-tone pixels. In order to provide a system giving high print quality and in which smoothing and enlarging operations are both carried out, the device is equipped with a facility (22) which determines for each two-tone pixel of the two-tone pixels, a copy pattern which matches a pixel arrangement having pixels which are arranged around each two-tone pixel. A device (23) converts the respective two-tone pixel either into a discrete multi-tone pixel or into a number of multi-tone pixels.

Image data processor for representing gray levels by a plurality of pixels

Abstract: Disclosed is an image data processor for processing input image data representing the gray level of each of pixels constituting an image to produce output image data representing the gray level of each of pixels constituting an output image. The relative position of the pixel corresponding to the input image data in a group of pixels including a predetermined number of pixels is identified. The input image data is subjected to data processing corresponding to the relative position of the pixel in the group of pixels on the basis of the result of the identification. Consequently, gray levels are represented by the entire group of pixels.

Fractal features of the digital measurement of length and area

Abstract: For objects viewed within images composed of an array of square pixels, the precise results of length and area measurements depend on the relative size of objects and pixels. This study concerns the error to be expected in such measurements. Not surprisingly, the log-log relationship between error and relative pixel size, i.e. relative size of the measurement unit, is linear with a non-integer gradient. A relationship of this form has been taken as the signature of a fractal phenomenon. However, such an interpretation of the images in question may not be of significant use in practical applications of quality control conducted by automated visual inspection. In the theoretical and experimental (computer simulation) studies reported here, each pixel has been represented by its centre point, lines have been represented by the sequence of four-connected pixels lying closest to the line and areas by the pixels of which the centres lie within the boundary of the figure being studied. Grey levels have not been used: each pixel is coloured either black or white. Pixels are taken as separated by unit distance and being of unit area. The restriction to one-bit pixel representation and four-connection between neighbouring pixels emphasizes the problems of geometric probability and discrete mathematics. Some small mitigation of the difficulties can be obtained by using multi- (grey) -coloured pixels and eight-connection between them. (3 pages)