Number Of Gray Levels Research Articles

Influence of the luminance spatial non-homogeneities of a display monitor on the contrast sensitivity function determination

The application of computer devices equipped with video monitors both in clinical practice and human vision research has become widespread. These devices offer great advantages over other traditional methodologies (such as printed tests) because they facilitate the variation of the experimental parameters or the design of new psychophysical experiences. The aim of this paper is to determine the degree to which spatial inhomogeneity affects the generation of sinusoidal gratings on a video display. Knowing this degree of tolerance will enable us to determine the minimum demands that can be made on the graphic board and the video screens for their implementation in assessing contrast sensitivity in clinical practice. The importance of contrast deviation—due to sampling and discrete number of gray levels—exceeds that of spatial inhomogeneities. Therefore, correct stimulus generation should not be expressed in location space but in frequency space, as then we will ensure both the stimulus spatial frequency and contrast.

Gray-scale data pages for digital holographic data storage

The prospects for gray-scale (or multilevel) digital holographic data storage are theoretically and experimentally investigated. A simple signal-to-noise ratio (SNR) partitioning argument shows that when SNR scales as 1 over the number of holograms squared, five gray levels (log(2) 5 bits/pixel) would be expected to result in a 15% capacity increase over binary data pages. However, the additional signal-dependent noise sources present in practical systems create a baseline SNR that reduces both the optimal number of gray levels and the resulting gain in capacity. To implement gray-scale recording experimentally, we adapt the predistortion technique previously developed for binary page-oriented memories [Opt. Lett. 23, 289 (1998)]. Several new block-based modulation codes for decoding gray-scale data pages are introduced. User capacity is evaluated by an experimental technique using LiNbO(3) :Fe in the 90 degrees geometry. Experimental results show that a balanced modulation code with three gray levels provides a 30% increase in capacity (as well as a 30% increase in readout rate) over local binary thresholding.

Time and space results of dynamic texture feature extraction in MR and CT image analysis.

Texture feature extraction is a fundamental part of texture image analysis. Therefore, the reduction of its computational time and storage requirements should be an aim of continuous research. The Spatial Grey Level Dependence Method (SGLDM) is one of the most important statistical texture description methods, especially in medical image analysis. Co-occurrence matrices are employed for the implementation of this method; however, they are inefficient in terms of computational time and memory space, due to their dependency on the number of gray levels (gray-level range) in the entire image. Since texture is usually measured in a small image region, a large amount of memory is wasted while the computational time of the texture feature extraction operations is unnecessarily raised. Their inefficiency puts up barriers to the wider utilization of SGLDM in a real application environment, such as a clinical environment. In this paper, the memory space and time efficiency of a dynamic approach to texture feature extraction in SGLDM is investigated through a pilot application in the analysis of magnetic resonance (MR) and computed tomography (CT) images.

Optical Flow Calculation Using Look-Up Tables

This work presents a new approach to optical flow calculation from an image sequence. Previously published methods were characterised by their high computational cost and complexity, limiting their applicability. This approach is characterised by avoiding floating point calculations or its emulation using integer arithmetic. Instead of this, starts with a previous off-line calculation of some LUT containing the resulting values of all the necessary operations for flow calculation. In the on-line phase of the method, the LUT is directly accessed using the gray-level of the point neighbourhood in one frame and the corresponding search space in the next frame. The vector formed by the gray levels of both neighbourhoods is interpreted as a number in a numerical base equal to image depth. Those numbers are the indexes to access the LUT and obtain the correct displacement value directly. In order to have some preliminary results, the test of the algorithm was performed with Matlab 5.0, and the results were compared with traditional correlation based methods. This implementation performed nine times faster than the simple correlation based algorithm. Despite of the reduction in the number of gray levels of the image, achieved precision is equal to the one from correlation. According to these results, near real-time optical flow computation can be achieved, with no special hardware requirements.

Window width as a dosage-relevant factor in high-contrast structures in CT

To establish the relationship between window width and dose to be applied in low-dose high-resolution (HR) computed tomography (CT). Low-dose HR CT-scans of the petrous bone displayed with different window values were analyzed for identification of osseous details. For two homogeneous phantoms, standard deviation of CT-numbers were measured in order to calculate the fraction of pixels not displayed within the correct grey level. The broadest window used (4000 HU) allowed the best distinction of osseous structures. Standard deviation of CT-numbers in the phantoms varied between 28.5 and 43.2 (mean = 33.6) HU. Thus, only 26.6% of all pixels are displayed in the correct grey level for a 1000 HU-window, but 82.3% for a 4000 HU-window. When using low doses and HR-reconstruction algorithms, large window widths allow for an optimal assessment of high-contrast structures. Under these conditions, even high standard deviations of the CT-numbers will not compromise image evaluation. Due to the restricted number of grey-levels that can be distinguished by the human eye, image noise is compressed into a smaller number of discernable grey-levels. Because of the inverse square root function between dose to be applied and window width used (derived in the article), an even minor enlargement of window permits considerable reductions of patient exposure while the discernable image noise remains constant.

Dual nonlinear correlator based on computer controlled joint transform processor: digital analysis and optical results

Abstract The hybrid optoelectronic processor, presented in this paper, realizes the dual nonlinear correlation (DNC) in a set-up based on a two-step nonlinear joint transform correlator architecture. In the first step three power spectrum distributions (input scene power spectrum, reference target power spectrum, and the joint power spectrum) necessary for the nonlinear processing are captured with a CCD camera. Nonlinear modification of the joint power spectrum, which does not have to be symmetrical in the input and reference channels, is introduced digitally. In the second step, the modified joint power spectrum is Fourier transformed optically. Numerical analysis of this processor shows a crucial influence of the dynamic range and the limited number of grey levels of the CCD camera during image acquisition in the first step, on the output signal parameters and the discrimination capability of the set-up. Optical results of recognition obtained for noise-free segmented input scenes show that the set-up enables the realization of various correlation operations as CMF, POF, IF, PPC, etc. without any filters and that the discrimination capability is easy to control by a proper selection of the type of nonlinear processing.

Dual nonlinear correlator based on computer controlled joint transform processor: Digital analysis and optical results

The hybrid optoelectronic processor, presented in this paper, realizes the dual nonlinear correlation (DNC) in a set-up based on a two-step nonlinear joint transform correlator architecture. In the first step three power spectrum distributions (input scene power spectrum, reference target power spectrum, and the joint power spectrum) necessary for the nonlinear processing are captured with a CCD camera. Nonlinear modification of the joint power spectrum, which does not have to be symmetrical in the input and reference channels, is introduced digitally. In the second step, the modified joint power spectrum is Fourier transformed optically. Numerical analysis of this processor shows a crucial influence of the dynamic range and the limited number of grey levels of the CCD camera during image acquisition in the first step, on the output signal parameters and the discrimination capability of the set-up. Optical results of recognition obtained for noise-free segmented input scenes show that the set-up enables the realization of various correlation operations as CMF, POF, IF, PPC, etc. without any filters and that the discrimination capability is easy to control by a proper selection of the type of nonlinear processing.

Effects of Image Quality, Number of Selectable Viewpoints, and Way to Select the Viewpoint in X-Ray Luggage Inspection

For luggage inspection, a good 3D impression is needed. This impression can be enhanced either by improving the image quality or by providing multiple viewpoints. An experiment is described in which students had to judge whether two objects in a transparent box were connected with a wire. Subjects using head movements to select the viewpoint were compared with subjects using a knob. Viewpoint selection was limited to horizontal movement. The resolution and the number of grey levels of the image as well as the size and the number of discrete steps of image movement were manipulated. Results show that 16 grey levels are sufficient to detect the wire, but that for a correct decision about connectivity, a freedom of movement of at least 90° is desirable. No difference was found between selecting the desired viewpoint with the head and by knob.

Thickness-equalization processing for mammographic images.

A digital postprocessing technique was used to compensate for the limitations of laser film or cathode-ray-tube devices used to display digital mammograms. An algorithm identified and equalized for the large change in digital signal caused by the reduction in thickness at the margin of the compressed breast. The resulting images reflected only breast composition, and so the number of gray levels needed to display the processed image was greatly reduced, which facilitated presentation and analysis.

Characterization of echographic image texture by cooccurrence matrix parameters

The goal of this study was to investigate the potentials of cooccurrence matrix analysis for the characterization of echographic image texture. Echographic data were obtained by one-dimensional simulations. Various data sets were generated with different number densities of the randomly distributed scatterers and with different levels of structural scattering strength. Cooccurrence matrix parameters estimated for analysis were: angular second moment, contrast, correlation, entropy and kappa. The cooccurrence matrix analysis was tuned by varying its parameters: spatial displacement of the pixel pairs, number of gray levels and size of the window. The parameters reached a saturation level at a number density of 3–5 scatterers per resolution cell (−6 dB width). Similarly, when increasing the displacement, a limit value was reached at 4–20 samples, depending on the size of the resolution cell. Using the Mahalanobis distance as a measure of differentiating between two textures, a systematic inverse relation was observed between the size of the window and the number of gray levels used in the estimation of the cooccurrence matrix. The optimal parameters to differentiate textures without structure appeared to be entropy and angular second moment. A window size of 90 speckle cells and 64 gray levels is needed for this purpose. The effective speckle size was estimated from the mean number of maxima of the demodulated echo signals. Resolved structure results in a periodicity of parameter values with displacement. The periodicity can be calculated by changing the displacement d. Optimal parameters for detecting periodicity are contrast and correlation. Analysis of the correlation between parameters showed that entropy vs. angular second moment and contrast vs. correlation are highly correlated.

CNN paradigm based multilevel halftoning of digital images

An algorithm for displaying gray level images using a small number of fixed quantization levels is proposed. The algorithm, called multilevel halftoning, is based on the Cellular Neural Networks (CNN) paradigm. It tracks the CNN transient outputs and selects the image which is subjectively perceived to be the best when reduced to the allowed number of gray levels. The selection criterion is based on the visually compensated mean square error that takes into account the specifics of the human visual system. The results of the proposed algorithm were validated in subjective quality experiments with human subjects.

Hierarchical fast two‐dimensional entropic thresholding algorithm using a histogram pyramid

Entropic thresholding provides an alternative view to the de- sign rationale of conventional thresholding. Abutaleb proposed a second-order entropic thresholding approach to improve Pun's first-order entropic thresholding by introducing a 2-D gray-level histogram to take into account the spatial correlation. Abutaleb's (1989) method was fur- ther modified by Brink (1992). However, a major drawback of Abutaleb- Brink's method is very high computational cost. Recently, Chen et al. (1994) developed a fast efficient 2-D algorithm that reduces computa- tional complexity from O(L 4 )t o O ( L 8/3 ), where L is the total number of gray levels. A hierarchical fast 2-D entropic thresholding algorithm using a gray-level histogram pyramid is presented that can be viewed as a generalization of Chen et al.'s algorithm. The new algorithm consists of a 2-D gray-level histogram pyramid build-up procedure expanding Ab- utaleb's 2-D gray-level histogram to a histogram pyramid, and a thresh- olding process applying a modified version of Chen et al.'s algorithm to the histogram pyramid layer by layer from top to bottom. As a result, the computational complexity of Chen et al.'s algorithm can be further re- duced to the optimal complexity, O(L 2 ). The experiments show that the computer time of the new algorithm is only one tenth of that required for Chen et al.'s algorithm, which is a significant saving. © 1996 Society of Photo-Optical Instrumentation Engineers.

Computational method for testing computer‐generated holograms

A procedure based entirely on computer simulation is developed for testing the quality of computer-generated holographic optical elements (CGHOEs). This new testing method may be of help in optimization techniques in the future. We introduce a mathematical quantity, the correlation, to characterize the performance of the studied CGHOE. Our procedure examines how the correlation between the object and the reconstructed image depends on various parameters of the CGHOE, such as the reference amplitude, the location and number of the object points, the number of gray levels used, and the pixel size on the hologram. To reduce computer running time, we introduce a new method for modeling the reconstruction of a computer-generated hologram (CGH): instead of using the whole information stored on the CGH, we use randomly chosen pixels (square-shaped slits) of the CGH. In this way, computer running time becomes smaller by two orders. A postquantization procedure is developed that, under certain conditions, improves the quality of the CGHOE. Our model uses amplitude holograms and 2-D objects, but in principle, there are no difficulties in extending it to phase holograms and 3-D objects.

Visible Speech as a Function of Image Quality: Effects of Display Parameters on Lipreading Ability

Two experiments are reported that examine effects of certain visual display parameters on subjects' ability to lipread numbers. The quality of the image of the speaker's face is altered by varying temporal resolution (frame rate), spatial resolution (number of pixels) and contrast resolution (number of grey scale levels). In Experiment 1, combinations of these variables are examined in a repeated measures design and reliable effects of reducing both frame rate and number of grey levels are evident. In Experiment 2, the frame rate and grey scale settings are factorially combined on a between-subjects basis, and practice is given so that subjects can adapt to the display conditions to which they are exposed. Effects of frame rate and grey scale are again statistically significant, and additive. Lipreading performance improves over the short duration of the experiment, but both image variables continue to affect performance. The effect size of the grey scale factor decreases with practice, but not that of frame rate. The results are discussed with regard to the kinds of relevant visual cues that may be affected by image degradation. Implications for the planned development of low cost digital communication are also considered. Trading off one image variable at the expense of another may be necessary to meet bandwidth limitations, and data such as those presented will be necessary to decide on a combination of parameter values to satisfy human performance and technical criteria.

Transformation of gray level and color images

Image transformations play a major role in image processing, pattern recognition, and computer vision. One of the applications of image transformation is to convert an input image into another one, which can be processed more easily and/or with a better result. Another application is to generate various new sample images from an existing image. Most of the papers dealing with image transformation describe algorithms that can only process binary images, and therefore have limited applications. In this paper, new image transformation algorithms capable of handling gray level and color images are proposed. These algorithms perform the mapping and filling at the same time, while preserving the connectivity of the original image. Also, they keep the continuity of the intensities by assigning the proper value of the neighbors to adequately fill the empty spots. The proposed algorithms can deal with all kinds of images, and are applicable to producing a large number of gray level and color images in a short time, hence, they will be very useful for image processing and pattern recognition.

Trade-off between resolution and interactivity in spatial task performance

Virtual reality displays usually lag far behind classical computer graphics displays in static image quality parameters, such as resolution. Both the popular press and scientific papers often stress that resolution will have to increase greatly before users can experience virtual environments as "the real thing". Nevertheless, it is already possible to do some useful work in VR environments. The point we experimentally demonstrate here is that resolution is much less important for interactive tasks that employ immersive VR, where users can explore the environment by moving their heads and bodies, than it is in classical computer graphics applications, where users can only explore by gazing at a single picture. In the context of unmanned aerial vehicles, frame rate (read: passive camera movement) is more important than resolution for target detection, recognition, designation, and tracking. In the experiments reported here, we investigated the relative importance of various image parameters like spatial resolution (number of pixels per video frame), intensity resolution (number of gray levels per pixel), and temporal resolution (number of frame updates per second). Most experimental data concerning these resolutions come from classical psychophysics. However, experimental conditions in classical psychophysics feature stationary observers looking at short-term, pointlike flashes on stationary displays, and are thus far more representative of human interaction with pictures and photographs than with highly interactive systems like those employed in virtual reality.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

The Video Camera Compared with the Densitometer as a Scanning Device for Microradiographγ

In the recent past image analysis systems, comprising a video (CCD) camera and dedicated software, have replaced densitometer-based systems to analyze mineral content profiles using transversal microradiography (TMR). The main reasons for the introduction of the CCD camera were the ease and speed at which it can be operated. The densitometer, as a scanning device for TMR, has in the recent years been validated and is in this study considered as 'gold standard'. Comparisons of the two scanning devices for measuring the optical density of microradiograms have never been reported in the literature. The focus of this study was on accuracy and reproducibility of the scanning devices with the emphasis put on possible limitations of the CCD camera relative to the densitometer. These include resolution, number of gray levels and homogeneity of illumination of the scan area. The microscope was arranged in such a way that the same area on the microradiogram could be assessed by both scanning devices. Three different types of lesions were analyzed: a subsurface lesion, a surface-softened lesion and a laminated lesion. Paired t tests showed no significant difference between the mineral content profiles produced by the two scanning devices. The integrated mineral loss values were calculated and compared with analysis of variance and showed no significant differences. It is therefore concluded that mineral content profiles and integrated mineral loss obtained by the CCD camera are as accurate and reproducible as those obtained by the densitometer.

Modelling micro-texture

There are many instances when texture contains valuable information in images, and various methods have been used for texture analysis. We distinguish between micro-textures and macro-textures. The paper models micro-texture using the general spin Ising model from statistical mechanics. This model allows for any number of grey levels and any set of pair interactions. For a given texture, we select an appropriate set of pair interactions and estimate the correspomding parameter values, using linked cluster expansions of the auto-covariances and the partition function. The series expansions are valid for parameters smaller than the critical parameters for which an infinite system would exhibit a phase transition. Hence, sufficiently small-grained micro-textures may be modelled. To ensure that the data meet this requirement, we simulate the model using the Markov chain Meet Carlo method and estimate its critical parameters using the series expansions. We demonstrate these methods on both real and simulated i...

Gibbs random fields, cooccurrences, and texture modeling

Gibbs random field (GRF) models and features from cooccurrence matrices are typically considered as separate but useful tools for texture discrimination. The authors show an explicit relationship between cooccurrences and a large class of GRF's. This result comes from a new framework based on a set-theoretic concept called the and on measures of this set, measures. This framework is also shown to be useful for relating different texture analysis tools. The authors show how the aura set can be constructed with morphological dilation, how its measure yields cooccurrences, and how it can be applied to characterizing the behavior of the Gibbs model for texture. In particular, they show how the aura measure generalizes, to any number of gray levels and neighborhood order, some properties previously known for just the binary, nearest-neighbor GRF. Finally, the authors illustrate how these properties can guide one's intuition about the types of GRF patterns which are most likely to form. >

Dynamic range requirements in digital mammography.

The dynamic range and the number of gray levels, gamma s, required for digital mammography has been evaluated using an energy transport model. The effects of molybdenum (Mo) and tungsten (W) target spectra and the energy-dependent attenuation by elemental filters, breast tissue, and a phosphor screen were included in the model. For detectors with ideal optical coupling and no inherent detector noise, 3,100 gray levels are discernable (requiring 12 bits per pixel), assuming a 40 kVp, W target spectrum (1.0 mm A1 filtration), a mean glandular dose to a 5 cm thick breast of 0.6 mGy, and an ideal observer with a 5 mm diam viewing aperture. The effects of inherent detector noise and realistic coupling efficiency on gamma s were also examined. For the 40 kVp, W spectrum, a detector with total coupling efficiency of 16 electrons (e-) per x-ray interaction and a dynamic range of 3000 (maximum carrier signal of 1.93 x 10(5) e-/pixel and inherent detector noise of 64 e- pixel) would decrease the number of gray levels that could be resolved by only 2% compared to a detector with ideal coupling and no inherent noise. A detector with a total coupling efficiency of 2.0 electrons per x-ray interaction and a dynamic range of 240 (maximum carrier signal 2.41 x 10(4) e-/pixel and inherent detector noise of 100 e-/pixel) would reduce the number of gray levels by 26% for the 40 kVp spectrum. On the basis of dynamic range, W spectra are preferable for digital mammography, since Mo spectra yielding the same signal-to-noise ratio require a detector with dynamic range twice as large, and with a 30% greater saturation signal. When no scatter rejection method is used, scattered radiation over a 254 cm2 imaging field reduces the number of discernable gray levels by 23% for a 5 cm thick breast and 34% for an 8 cm thick breast.