Showing papers on "Standard test image published in 2002"

Recognizing imprecisely localized, partially occluded, and expression variant faces from a single sample per class

Abstract: The classical way of attempting to solve the face (or object) recognition problem is by using large and representative data sets. In many applications, though, only one sample per class is available to the system. In this contribution, we describe a probabilistic approach that is able to compensate for imprecisely localized, partially occluded, and expression-variant faces even when only one single training sample per class is available to the system. To solve the localization problem, we find the subspace (within the feature space, e.g., eigenspace) that represents this error for each of the training images. To resolve the occlusion problem, each face is divided into k local regions which are analyzed in isolation. In contrast with other approaches where a simple voting space is used, we present a probabilistic method that analyzes how "good" a local match is. To make the recognition system less sensitive to the differences between the facial expression displayed on the training and the testing images, we weight the results obtained on each local area on the basis of how much of this local area is affected by the expression displayed on the current test image.

Speckle removal from SAR images in the undecimated wavelet domain

Abstract: Speckle reduction is approached as a minimum mean-square error (MMSE) filtering performed in the undecimated wavelet domain by means of an adaptive rescaling of the detail coefficients, whose amplitude is divided by the variance ratio of the noisy coefficient to the noise-free one. All the above quantities are analytically calculated from the speckled image, the variance and autocorrelation of the fading variable, and the wavelet filters only, without resorting to any model to describe the underlying backscatter. On the test image Lena corrupted by synthetic speckle, the proposed method outperforms Kuan's local linear MMSE filtering by almost 3-dB signal-to-noise ratio. When true synthetic aperture radar (SAR) images are concerned, empirical criteria based on distributions of multiscale local coefficient of variation, calculated in the undecimated wavelet domain, are introduced to mitigate the rescaling of coefficients in highly heterogeneous areas where the speckle does not obey a fully developed model, to avoid blurring strong textures and point targets. Experiments carried out on widespread test SAR images and on a speckled mosaic image, comprising synthetic shapes, textures, and details from optical images, demonstrate that the visual quality of the results is excellent in terms of both background smoothing and preservation of edge sharpness, textures, and point targets. The absence of decimation in the wavelet decomposition avoids typical impairments often introduced by critically subsampled wavelet-based denoising.

Object Recognition using Local Affine Frames on Distinguished Regions

Abstract: A novel approach to appearance based object recognition is introduced. The proposed method, based on matching of local image features, reliably recognises objects under very different viewing conditions. First, distinguished regions of data-dependent shape are robustly detected. On these regions, local affine frames are established using several affine invariant constructions. Direct comparison of photometrically normalised colour intensities in local, geometrically aligned frames results in a matching scheme that is invariant to piecewise-affine image deformations, but still remains very discriminative. The potential of the approach is experimentally verified on COIL-100 and SOIL-47 ‐ publicly available image databases. On SOIL-47, 100% recognition rate is achieved for single training view per object. On COIL-100, 99.9% recognition rate is obtained for 18 training views per object. Robustness to severe occlusions is demonstrated by only a moderate decrease of recognition performance in an experiment where half of each test image is erased.

Combined image views and method of creating images

Abstract: A first image of a first software which can be combined with other images of other software such that any one or more of the images can be restored from the combined image, and methods relating thereto. The method of making the combined image comprises creating a first image from a first software, creating a second image from the second software, and combining the first image and the second image into the combined image. Each image includes first descriptive data (metadata) corresponding to descriptive data of its software and includes file data corresponding to file data of its software.

Recognizing handwritten digits using hierarchical products of experts

Abstract: The product of experts learning procedure can discover a set of stochastic binary features that constitute a nonlinear generative model of handwritten images of digits. The quality of generative models learned in this way can be assessed by learning a separate model for each class of digit and then comparing the unnormalized probabilities of test images under the 10 different class-specific models. To improve discriminative performance, a hierarchy of separate models can be learned, for each digit class. Each model in the hierarchy learns a layer of binary feature detectors that model the probability distribution of vectors of activity of feature detectors in the layer below. The models in the hierarchy are trained sequentially and each model uses a layer of binary feature detectors to learn a generative model of the patterns of feature activities in the preceding layer. After training, each layer of feature detectors produces a separate, unnormalized log probability score. With three layers of feature detectors for each of the 10 digit classes, a test image produces 30 scores which can be used as inputs to a supervised, logistic classification network that is trained on separate data.

Method for computing visual performance from objective ocular aberration measurements

Abstract: A method for computing the visual performance of a human or animal subject based on objective measurements of visual refraction, including higher order aberrations, includes measuring wavefront aberrations of a subject ocular pupil, computing a point-spread-function from the measured pupil aberration, providing a test image, and convolving the test image with the point-spread-function. A simulated image may be produced from the convolution result of the test image with the point-spread-function. One or more specific terms of the point-spread-function may be subtracted therefrom prior to the convolving step to simulate an effect of a correcting means, such as spectacles lenses, contact lenses, or laser surgery. A best correction for a given subject may be determined by adjusting the terms that are subtracted to optimize the resultant image.

Image processing method, image processing apparatus and program therefor

Abstract: Predetermined processing such as image processing is carried out on image data without carrying out a troublesome operation. A folder F in a hard disc 2 of a personal computer 1 is related to processing for calculating image processing parameters P used in image processing on image data sets S input by an image processing daemon, for carrying out the image processing according to the image processing parameters P, and for storing processed image data sets S1 obtained by the image processing. The image processing daemon carries out the processing in the background of the operating system, when the image data sets S are input to the folder F, and obtains the processed image data sets S1.

3-D face structure extraction and recognition from images using 3-D morphing and distance mapping

Abstract: We describe a novel approach for creating a three-dimensional (3-D) face structure from multiple image views of a human face taken at a priori unknown poses by appropriately morphing a generic 3-D face. A cubic explicit polynomial in 3-D is used to morph a generic face into the specific face structure. The 3-D face structure allows for accurate pose estimation as well as the synthesis of virtual images to be matched with a test image for face identification. The estimation of a 3-D person's face and pose estimation is achieved through the use of a distance map metric. This distance map residual error (geometric-based face classifier) and the image intensity residual error are fused in identifying a person in the database from one or more arbitrary image view(s). Experimental results are shown on simulated data in the presence of noise, as well as for images of real faces, and promising results are obtained.

Image processing and enhancement provided by commercial dental software programs.

Abstract: OBJECTIVES To identify and analyse methods/algorithms for image processing provided by various commercial software programs used in direct digital dental imaging and to map them onto a standardized nomenclature. METHODS Twelve programs presented at the 28th International Dental-Show, March, 2001, Cologne, Germany and the Emago advanced software were included in this study. An artificial test image, comprised of gray scale ramps, step wedges, fields with Gaussian-distributed noise, and salt and pepper noise, was synthesized and imported to all programs to classify algorithms for display; linear, non-linear and histogram-based point processing; pseudo-coloration; linear and non-linear spatial filtering; frequency domain filtering; measurements; image analysis; and annotations. RESULTS The 13 programs were found to possess a great variety of image processing and enhancement facilities. All programs offer gray-scale image display with interactive brightness and contrast adjustment and gray-scale inversion as ...

Digital image compression

Abstract: A method of reconstructing an image, where the input image data is preferably part I or part II compliant JPEG2000 coded data, or pixel data of the original image. The method selects ( 810 ) an output resolution R, and then determines a number of sub-passes to extract from each block code based on the selected resolution. The method then extracts ( 830 ) the determined sub-passes and the remaining sub-passes are discarded. The method then reconstructs 840 the image from the extracted sub-passes. The reconstructed image can be in the form of the selected resolution of the original image, or it can be in the form of compressed image data of the selected resolution of the original image.

Statistical analysis of nonlinearly reconstructed near-infrared tomographic images. I. Theory and simulations

Abstract: Near-infrared (NIR) diffuse tomography is an emerging method for imaging the interior of tissues to quantify concentrations of hemoglobin and exogenous chromophores noninvasively in vivo. It often exploits an optical diffusion model-based image reconstruction algorithm to estimate spatial property values from measurements of the light flux at the surface of the tissue. In this study, mean-squared error (MSE) over the image is used to evaluate methods for regularizing the ill-posed inverse image reconstruction problem in NIR tomography. Estimates of image bias and image standard deviation were calculated based upon 100 repeated reconstructions of a test image with randomly distributed noise added to the light flux measurements. It was observed that the bias error dominates at high regularization parameter values while variance dominates as the algorithm is allowed to approach the optimal solution. This optimum does not necessarily correspond to the minimum projection error solution, but typically requires further iteration with a decreasing regularization parameter to reach the lowest image error. Increasing measurement noise causes a need to constrain the minimum regularization parameter to higher values in order to achieve a minimum in the overall image MSE.

Automatic tone mapping for images

Abstract: An image processing method for processing image data comprises the steps of obtaining scanpath data corresponding to original image data, determining regions of interest for the original image data based on the obtained scanpath data, and mapping tone values of the original image data corresponding to each region of interest in order to obtain tone-mapped image data.

Method and apparatus for image improvement

Abstract: A method and an apparatus for improving the image quality. The method comprises steps of producing image correcting information to decrease errors in the image to be produced by a camera module (109, 111, 112) by comparing the image taken by said camera module (109, 111, 112, 113) with a test image, storing the image correcting information produced and identifying information related to said image correcting information in the memory of an electronic peripheral device (104, 106), receiving the image produced by said camera module (109, 111, 112, 113) and a second piece of identifying information related to said image in the electronic peripheral device (104, 106), comparing said identifying information with said second piece of identifying information in said peripheral device, and performing an image improvement operation on said image in said peripheral device in response to the comparison carried out.

Image retrieval method based on combination of color and texture features

Abstract: An image retrieval method with improved performance in which a data image similar to a query image is retrieved by appropriately combining color and texture features. The method for retrieving a data image similar to a query image in an image database containing a plurality of data images involves: calculating color and texture distances between a query image and each data image in the image database; weighting the calculated color and texture distances with predetermined weighting factors; calculating a feature distance between the query image and each data image by combining the weighted color and texture distances by considering human visual perception attributes; and determining the data image similar to the query image using the feature distance. According to the image retrieval method, data images similar to a query image can be retrieved based on the human visual perception mechanism by combining both color and texture features extracted from image regions. In particular, the image region based retrieval enables accurate retrieval of more objects and many kinds of information from a single image.

A Five-Step Approach to Digital Image Manipulation for the Radiologist

Abstract: Digital manipulation of images plays a key role in development of multimedia presentations. Five basic steps to digitizing images and preparing them for publication and computer presentation are scanning, correction, editing and labeling, saving files, and producing final output. These steps can be completed with commercially available hardware and image manipulation software (eg, Photoshop). The higher the quality of the original scanned image, the more image data there will be to edit: A good image cannot be created from an inferior scan. The most important functions for properly scanning images are size, resolution, and color. Resolutions of 300 ppi and 72 ppi should be used for print publication and computer presentations, respectively. The higher resolution image has the larger file size. The scanned image should be saved as a TIFF (tagged image file format), which is an uncompressed file type used for printed images. The Joint Photographic Experts Group (JPEG) format compresses the size of the image file but also reduces image quality. The JPEG format is a good choice if a small file size is needed, such as in Web and PowerPoint presentations. If the user needs to save an image as a JPEG file, the image should be edited first and then saved once in JPEG format. With Photoshop, the user can rotate and crop an image; adjust its brightness, contrast, and color; remove unwanted patient information, dust, and scratches; and add text and symbol labels to enhance images for teaching purposes. Digital manipulation can be fast and effective if the user has some basic knowledge and tools.

Method and apparatus for global image quantification verification

Abstract: A method and apparatus are disclosed for testing the accuracy of digital test images generated by a computer graphics program executed on a computer graphics system. A test program is utilized to compare the test image with a reference image. The user defines regions of interest and a maximum error for each region. This allows the user to focus on a particular object or group of objects. Global image quantification verification generates one measurement of error or image difference for each region of the test image. Each region is divided into sub-regions and an average color value is calculated for each sub-region and the corresponding sub-region in the reference image. The differences between the averages for corresponding sub-regions in the test and reference images are calculated. A test image region may be unacceptably different from a reference image if a normalized sum of the absolute differences exceeds the maximum error.

Noise reduction techniques for Bayer-matrix images

Abstract: In this paper, some arrangements to apply Noise Reduction (NR) techniques for images captured by a single sensor digital camera are studied. Usually, the NR filter processes full three-color component image data. This requires that raw Bayer-matrix image data, available from the image sensor, is first interpolated by using Color Filter Array Interpolation (CFAI) method. Another choice is that the raw Bayer-matrix image data is processed directly. The advantages and disadvantages of both processing orders, before (pre-) CFAI and after (post-) CFAI, are studied with linear, multi-stage median, multistage median hybrid and median-rational filters .The comparison is based on the quality of the output image, the processing power requirements and the amount of memory needed. Also the solution, which improves preservation of details in the NR filtering before the CFAI, is proposed.

Automatic image quality evaluation and correction technique for digitized and thresholded document images

Abstract: A process and system for automating the image quality inspection and correction for scanned document which previously required a human operator. For every scanned and thresholded image, the process and system performs an automatic evaluation through a binary image quality detection system which generates an image noise index indicative of the amount of image artifacts or image loss. When a poor quality scanned page is detected, for example, too much speckle noise, the gray scale image is retrieved or the image is rescanned. The gray scale image then automatically undergoes an image quality correction process to produce a clean, readable binary image.

Text scanner with text detection technology on image sequences

Abstract: We propose a text scanner which detects wide text strings in a sequence of scene images. For scene text detection, we use a multiple-CAMShift algorithm on a text probability image produced by a multi-layer perceptron. To provide enhanced resolution of the extracted text images, we perform the text detection process after generating a mosaic image in a fast and robust image registration method.

Protective device for safeguarding a hazardous area and method of checking the functional reliability of such a device

Abstract: The present invention discloses a protective device for safeguarding a hazardous area and a method for checking the functional reliability of such a device. The protective device has an image recording unit with an image sensor which includes a large number of light-sensitive pixels. During operation, the image recording unit records an object image. The image sensor is also supplied with a defined test image in order to check the functional reliability, wherein the test image recorded by the image sensor is compared with a defined expectation. According to one aspect of the invention, the object image is specifically made dynamic by means of a testing device and the modified object image is used as the test image.

Image data processing method, image data processing program, and stereoscopic image display apparatus

Abstract: PROBLEM TO BE SOLVED: To provide an image processing method and an image data processing program that allow various stereoscopic image display systems to share stereoscopic image data through the capability of converting the image data of the various stereoscopic image display systems into a particular system. SOLUTION: The image processing method and the image data processing program whereby different stereoscopic image display systems mutually convert the stereoscopic image data, include the steps of: reading and storing stereoscopic image data SD of a conversion source; discriminating a stereoscopic image display system of the stored stereoscopic image data of the conversion source; separately extracting left eye image data DL and right eye image data DR from the stored stereoscopic image data of the conversion source; interpolating or interleaving pixel data from the left eye image data and the right eye image data to revise the vertical resolution and / or the horizontal resolution in order to adapt the extracted left eye image data and right eye image data to a stereoscopic image display system of a conversion destination; and by outputting the left eye image data and right eye image data, its resolution is improved altogether as new stereoscopic image data. COPYRIGHT: (C)2004,JPO

System and method for efficient compression of raster image data

Abstract: A system and method for compressing raster image data that efficiently processes data that contains the same value for each pixel. Images are compressed according to the Joint Photographic Experts Groups (JPEG) standard. Raster image data for an image or single color plane is analyzed and if the image is determined to contain the same value for each pixel, the processing produces and replicates pre-computed compressed data output segments that replicate the output of JPEG compression.

On modeling variations for face authentication

Abstract: We present a scheme for face authentication in the presence of variations. To deal with variations, such as facial expressions and registration errors, with which traditional appearance-based methods do not perform well, we propose the eigenflow approach. In this approach, the optical flow and the optical flow residue between a test image and a training image are computed first. The optical flow is then fitted to a model that is pre-trained by applying principal component analysis (PCA) to optical flows resulting from variations caused by facial expressions and registration errors. The eigenflow residue, optimally combined with the optical flow residue using linear discriminant analysis (LDA), determines the authenticity of the test image. Experimental results show that the proposed scheme outperforms the traditional methods in the presence of expression variations and registration errors. The approach can be extended to model lighting and pose variations as well.

A survey and evaluation of edge detection operators application to medical images

Abstract: One of the objectives of image analysis is to extract its dominating information. Thus we use segmentation to associate a stamp to each pixel according to the carried information (gray level or color) and its specific distribution in the image. Thereby, the segmentation of the image is defined as being the low level step of processing that extracts and describes present significant objects in a scene, the most often in the form of regions or edges. In the literature, different methods have been elaborated in order to detect image edges. They are gathered in two families: on the one hand methods privileging an approach by border (derivative, surfaces, and morphological methods) named the edge approach; on the other hand those privileging an approach by regions (Markovian and structural methods). In this work, we are interested in the different methods using the edge approach for the image segmentation. Many image segmentation techniques are available. We describe derivative methods, optimal filtering, and segmentation for color images.

Method and apparatus for local image quantification verification

Abstract: A method and apparatus are disclosed for testing the accuracy of digital test images by comparing the value of each test image pixel with the average value of a corresponding reference image pixel array. The local image quantification verification test program accepts absolute differences between corresponding pixel and array values that are less than a specified difference maximum. The user may specify a difference maximum and a maximum array size, and may restrict the test image comparison to specified regions of the test image with specific difference tolerances for each region. This allows the user to focus the comparison on certain regions as more important than others and to focus a region on a particular object or group of objects.

Method and apparatus for displaying arbitrarily magnified high resolution images using compressed domain processing

Abstract: A method for efficiently processing image data for display on a computer monitor is provided. The method initiates with reading image data in a compressed format into a memory associated with the computer. Then, at least one scaled copy of the image data is generated in the compressed format. Next, a display scale is determined for an image to be presented on a computer monitor. If the magnification is less than 100% of an image scale associated with the image data, then, either the image data or the at least one scaled copy of the image data is determined as being closest in magnification to the display scale, without being less than the display scale, to provide a closest in magnification image data. Next, the closest in magnification image data in the compressed format is processed. Then, the processed image data is displayed. The method is capable of displaying a portion of the image at any magnification. An integrated circuit chip and a system for efficiently processing image data for display are also provided.

Generation of image file

Abstract: In an image output system, image data generated by a digital still camera is output with a color printer. The digital still camera stores image processing control data preset by a maker of the digital still camera to be changeable for each image, and a user's setting of offset data. The image output system then generates an image file including the image data, the image processing control data, and the offset data. The color printer makes the image data to be subjected to a preset series of image processing based on the image processing control data (including the offset data) in the image file, and outputs a resulting processed image.