Showing papers in "electronic imaging in 2007"

The Blur Effect: Perception and Estimation with a New No-Reference Perceptual Blur Metric

Abstract: To achieve the best image quality, noise and artifacts are generally removed at the cost of a loss of details generating the blur effect. To control and quantify the emergence of the blur effect, blur metrics have already been proposed in the literature. By associating the blur effect with the edge spreading, these metrics are sensitive not only to the threshold choice to classify the edge, but also to the presence of noise which can mislead the edge detection. Based on the observation that we have difficulties to perceive differences between a blurred image and the same reblurred image, we propose a new approach which is not based on transient characteristics but on the discrimination between different levels of blur perceptible on the same picture. Using subjective tests and psychophysics functions, we validate our blur perception theory for a set of pictures which are naturally unsharp or more or less blurred through one or two-dimensional low-pass filters. Those tests show the robustness and the ability of the metric to evaluate not only the blur introduced by a restoration processing but also focal blur or motion blur. Requiring no reference and a low cost implementation, this new perceptual blur metric is applicable in a large domain from a simple metric to a means to fine-tune artifacts corrections.

Visual discomfort in stereoscopic displays : a review

Abstract: Visual discomfort has been the subject of considerable research in relation to stereoscopic and autostereoscopic displays, but remains an ambiguous concept used to denote a variety of subjective symptoms potentially related to different underlying processes. In this paper we clarify the importance of various causes and aspects of visual comfort. Classical causative factors such as excessive binocular parallax and accommodation-convergence conflict appear to be of minor importance when disparity values do not surpass one degree limit of visual angle, which still provides sufficient range to allow for satisfactory depth perception in consumer applications, such as stereoscopic television. Visual discomfort, however, may still occur within this limit and we believe the following factors to be the most pertinent in contributing to this: (1) excessive demand of accommodation-convergence linkage, e.g., by fast motion in depth, viewed at short distances, (2) 3D artefacts resulting from insufficient depth information in the incoming data signal yielding spatial and temporal inconsistencies, and (3) unnatural amounts of blur. In order to adequately characterize and understand visual discomfort, multiple types of measurements, both objective and subjective, are needed.

The smashed filter for compressive classification and target recognition

Abstract: The theory of compressive sensing (CS) enables the reconstruction of a sparse or compressible image or signal from a small set of linear, non-adaptive (even random) projections. However, in many applications, including object and target recognition, we are ultimately interested in making a decision about an image rather than computing a reconstruction. We propose here a framework for compressive classification that operates directly on the compressive measurements without first reconstructing the image. We dub the resulting dimensionally reduced matched filter the smashed filter. The first part of the theory maps traditional maximum likelihood hypothesis testing into the compressive domain; we find that the number of measurements required for a given classification performance level does not depend on the sparsity or compressibility of the images but only on the noise level. The second part of the theory applies the generalized maximum likelihood method to deal with unknown transformations such as the translation, scale, or viewing angle of a target object. We exploit the fact the set of transformed images forms a low-dimensional, nonlinear manifold in the high-dimensional image space. We find that the number of measurements required for a given classification performance level grows linearly in the dimensionality of the manifold but only logarithmically in the number of pixels/samples and image classes. Using both simulations and measurements from a new single-pixel compressive camera, we demonstrate the effectiveness of the smashed filter for target classification using very few measurements.

The usage of digital cameras as luminance meters

Abstract: Many luminance measuring tasks require a luminance distribution of the total viewing field. The approach of imageresolving luminance measurement, which could benefit from the continual development of position-resolving radiation detectors, represents a simplification of such measuring tasks. Luminance measure cameras already exist which are specially manufactured for measuring tasks with very high requirements. Due to high-precision solutions these cameras are very expensive and are not commercially viable for many image-resolving measuring tasks. Therefore, it is desirable to measure luminance with digital still cameras which are freely available at reasonable prices. This paper presents a method for the usage of digital still cameras as luminance meters independent of the exposure settings. A calibration of the camera is performed with the help of an OECF (opto-electronic conversion function) measurement and the luminance is calculated with the camera's digital RGB output values. The test method and computation of the luminance value irrespective of exposure variations is described. The error sources which influence the result of the luminance measurement are also discussed.

Vertical surround parallax correction

Abstract: Three general design for reducing parallax in a moving picture are disclosed, one design comprises stretching graphics content (fig. 3B) near the left (fig. 3B, 307) and right frame edges (fig. 3B, 308) of stereo pair (fig. 3B) elements in order to modify parallax settings. A second design comprises identifying a plurality of substantially corresponding points in a left eye view and right eye view (fig. 3B) and bringing the plurality of substantially corresponding points closer together using a morph technique, the third design entails fading a portion of a right eye view with a portion of a left eye view near or proximate an edge of the image.

Tone mapping for high-dynamic range displays

Abstract: We address the problem of re-rendering images to high dynamic range (HDR) displays, which were originally tone-mapped to standard displays. As these new HDR displays have a much larger dynamic range than standard displays, an image rendered to standard monitors is likely to look too bright when displayed on a HDR monitor. Moreover, because of the operations performed during capture and rendering to standard displays, the specular highlights are likely to have been clipped or compressed, which causes a loss of realism. We propose a tone scale function to re-render images first tone-mapped to standard displays, that focuses on the representation of specular highlights. The shape of the tone scale function depends on the segmentation of the input image into its diffuse and specular components. In this article, we describe a method to perform this segmentation automatically. Our method detects specular highlights by using two low-pass filters of different sizes combined with morphological operators. The results show that our method successfully detects small and middle sized specular highlights. The locations of specular highlights define a mask used for the construction of the tone scale function. We then propose two ways of applying the tone scale, the global version that applies the same curve to each pixel in the image and the local version that uses spatial information given by the mask to apply the tone scale differently to diffuse and to specular pixels.

Why do commodity graphics hardware boards (GPUs) work so well for acceleration of computed tomography

Abstract: Commodity graphics hardware boards (GPUs) have achieved remarkable speedups in various sub-areas of Computed Tomography (CT). This paper takes a close look at the GPU architecture and its programming model and describes a successful acceleration of Feldkamp's cone-beam CT reconstruction algorithm. Further, we will also have a comparative look at the new emerging Cell architecture in this regard, which similar to GPUs has also seen its first deployment in gaming and entertainment. To complete the discussion on high-performance PC-based computing platforms, we will also compare GPUs with FPGA (Field Programmable Gate Array) based medical imaging solutions.

High Dynamic Range Imaging Pipeline : Perception-Motivated Representation of Visual Content

Abstract: The advances in high dynamic range (HDR) imaging, especially in the display and camera technology, have a significant impact on the existing imaging systems. The assumptions of the traditional low-dynamic range imaging, designed for paper print as a major output medium, are ill suited for the range of visual material that is shown on modern displays. For example, the common assumption that the brightest color in an image is white can be hardly justified for high contrast LCD displays, not to mention next generation HDR displays, that can easily create bright highlights and the impression of self-luminous colors. We argue that high dynamic range representation can encode images regardless of the technology used to create and display them, with the accuracy that is only constrained by the limitations of the human eye and not a particular output medium. To facilitate the research on high dynamic range imaging, we have created a software package (http://pfstools.sourceforge.net/) capable of handling HDR data on all stages of image and video processing. The software package is available as open source under the General Public License and includes solutions for high quality image acquisition from multiple exposures, a range of tone mapping algorithms and a visual difference predictor for HDR images. Examples of shell scripts demonstrate how the software can be used for processing single images as well as video sequences.

Digital Camera Resolution Measurement Using Sinusoidal Siemens Stars

Abstract: The resolution of a digital camera is defined as its ability to reproduce fine detail in an image. To test this ability methods like the Slanted Edge SFR measurement developed by Burns and Williams1 and standardized in ISO 122332 are used. Since this method is - in terms of resolution measurements - only applicable to unsharpened and uncompressed data an additional method described in this paper had to be developed. This method is based on a Sinusoidal Siemens Star which is evaluated on a radius by radius or frequency by frequency basis. For the evaluation a freely available runtime program developed in MATLAB is used which creates the MTF of a camera system as the contrast over the frequency.

Image interpolation using multiscale geometric representations

Abstract: With the ever increasing computational power of modern day processors, it has become feasible to use more robust and computationally complex algorithms that increase the resolution of images without distorting edges and contours. We present a novel image interpolation algorithm that uses the new contourlet transform to improve the regularity of object boundaries in the generated images. By using a simple wavelet-based linear interpolation scheme as our initial estimate, we use an iterative projection process based on two constraints to drive our solution towards an improved high-resolution image. Our experimental results show that our new algorithm significantly outperforms linear interpolation in subjective quality, and in most cases, in terms of PSNR as well.

Codesign toolset for application-specific instruction-set processors

Abstract: Application-specific programmable processors tailored for the requirements at hand are often at the center of today's embedded systems. Therefore, it is not surprising that considerable effort has been spent on constructing tools that assist in codesigning application-specific processors for embedded systems. It is desirable that such design toolsets support an automated design flow from application source code down to synthesizable processor description and optimized machine code. In this paper, such a toolset is described. The toolset is based on a customizable processor architecture template, which is VLIW-derived architecture paradigm called Transport Triggered Architecture (TTA). The toolset addresses some of the pressing shortcomings found in existing toolsets, such as lack of automated exploration of the "design space", limited run time retargetability of the design tools or restrictions in the customization of the target processors.

3D body scanning technology for fashion and apparel industry

Abstract: This paper presents an overview of 3D body scanning technologies with applications to the fashion and apparel industry. Complete systems for the digitization of the human body exist since more than fifteen years. One of the main users of this technology with application in the textile field was the military industry. In fact, body scanning technology is being successfully employed since many years in military bases for a fast selection of the correct size of uniforms for the entire staff. Complete solutions were especially developed for this field of application. Many different research projects were issued for the exploitation of the same technology in the commercial field. Experiments were performed and start-up projects are to time running in different parts of the world by installing full body scanning systems in various locations such as shopping malls, boutiques or dedicated scanning centers. Everything is actually ready to be exploited and all the required hardware, software and solutions are available: full body scanning systems, software for the automatic and reliable extraction of body measurements, e-kiosk and web solutions for the presentation of garments, high-end and low-end virtual-try-on systems. However, complete solutions in this area have still not yet found the expected commercial success. Today, with the on-going large cost reduction given by the appearance of new competitors, methods for digitization of the human body becomes more interesting for the fashion and apparel industry. Therefore, a large expansion of these technologies is expected in the near future. To date, different methods are used commercially for the measurement of the human body. These can be divided into three major distinguished groups: laser-scanning, projection of light patterns, combination modeling and image processing. The different solutions have strengths and weaknesses that profile their suitability for specific applications. This paper gives an overview of their differences and characteristics and expresses clues for the selection of the adequate method. A special interest is given to practical examples of the commercial exploitation of human body digitization with applications to the fashion and apparel industry.

Autofocus survey: a comparison of algorithms

Abstract: There are numerous passive contrast sensing autofocus algorithms that are well documented in literature, but some aspects of their comparative performance have not been widely researched. This study explores the relative merits of a set of autofocus algorithms via examining them against a variety of scene conditions. We create a statistics engine that considers a scene taken through a range of focal values and then computes the best focal position using each autofocus algorithm. The process is repeated across a survey of test scenes containing different representative conditions. The results are assessed against focal positions which are determined by manually focusing the scenes. Through examining these results, we then derive conclusions about the relative merits of each autofocus algorithm with respect to the criteria accuracy and unimodality. Our study concludes that the basic 2D spatial gradient measurement approaches yield the best autofocus results in terms of accuracy and unimodality.

Experienced quality factors: qualitative evaluation approach to audiovisual quality

Abstract: Subjective evaluation is used to identify impairment factors of multimedia quality. The final quality is often formulated via quantitative experiments, but this approach has its constraints, as subject's quality interpretations, experiences and quality evaluation criteria are disregarded. To identify these quality evaluation factors, this study examined qualitatively the criteria participants used to evaluate audiovisual video quality. A semi-structured interview was conducted with 60 participants after a subjective audiovisual quality evaluation experiment. The assessment compared several, relatively low audio-video bitrate ratios with five different television contents on mobile device. In the analysis, methodological triangulation (grounded theory, Bayesian networks and correspondence analysis) was applied to approach the qualitative quality. The results showed that the most important evaluation criteria were the factors of visual quality, contents, factors of audio quality, usefulness - followability and audiovisual interaction. Several relations between the quality factors and the similarities between the contents were identified. As a research methodological recommendation, the focus on content and usage related factors need to be further examined to improve the quality evaluation experiments.

A Prototype 3D Mobile Phone Equipped with a Next Generation Autostereoscopic Display

Abstract: The mobile phone is quickly evolving from a communications device to a multimedia platform and in the process has become the focus for the development of new technologies. There is significant interest from mobile handset manufacturers as well as major telecommunications network operators in the development of a stereoscopic 3D mobile phone. The most challenging technical issues for commercializing a 3D phone are a stereoscopic display technology which is suitable for mobile applications as well as a means for driving the display using the limited capabilities of a mobile handset. In this paper we describe a prototype 3D mobile phone which was developed on a commercially available mobile hardware platform. The demonstration handset was retrofitted with a Polarization Activated Microlens™ array that is 2D/3D switchable and provides both class-leading low crosstalk levels, and suitable brightness characteristics and viewing zones for operation without compromising battery running time. This next generation autostereoscopic display technology, which combines the advantages in brightness and image quality of a lenticular 3D display with the 2D/3D switching capability of a liquid crystal parallax barrier, is deployed on a 2.2’’ landscape QVGA TFT-LCD base panel. The stereoscopic content solution is an essential component of a commercially viable 3D handset. We describe how a range of stereoscopic software solutions have been developed on the phone’s existing application processor without the need for custom hardware.

Bilinear Models of Natural Images

Abstract: Previous work on unsupervised learning has shown that it is possible to learn Gabor-like feature representations, similar to those employed in the primary visual cortex, from the statistics of natural images. However, such representations are still not readily suited for object recognition or other high-level visual tasks because they can change drastically as the image changes to due object motion, variations in viewpoint, lighting, and other factors. In this paper, we describe how bilinear image models can be used to learn independent representations of the invariances, and their transformations, in natural image sequences. These models provide the foundation for learning higher-order feature representations that could serve as models of higher stages of processing in the cortex, in addition to having practical merit for computer vision tasks.

High-dynamic-range video for photometric measurement of illumination

Abstract: We describe the design and implementation of a high dynamic range (HDR) imaging system capable of capturing RGB color images with a dynamic range of 10,000,000 : 1 at 25 frames per second. We use a ...

Multiscale reconstruction for computational spectral imaging

Abstract: In this work we develop a spectral imaging system and associated reconstruction methods that have been designed to exploit the theory of compressive sensing. Recent work in this emerging field indicates that when the signal of interest is very sparse (i.e. zero-valued at most locations) or highly compressible in some basis, relatively few incoherent observations are necessary to reconstruct the most significant non-zero signal components. Conventionally, spectral imaging systems measure complete data cubes and are subject to performance limiting tradeoffs between spectral and spatial resolution. We achieve single-shot full 3D data cube estimates by using compressed sensing reconstruction methods to process observations collected using an innovative, real-time, dual-disperser spectral imager. The physical system contains a transmissive coding element located between a pair of matched dispersers, so that each pixel measurement is the coded projection of the spectrum in the corresponding spatial location in the spectral data cube. Using a novel multiscale representation of the spectral image data cube, we are able to accurately reconstruct 256×256×15 spectral image cubes using just 256×256 measurements.

MOEVision: Simple Multiview Display with Clear Floating Image

Abstract: This paper presents a simple and inexpensive multiview 3D display system composed of a LCD panel, a convex lens array, and a Fresnel lens. In the proposed system a pair of the LCD fragment and a convex lens in the array plays the role of a projector. The idea of multiview 3D displays composed of multiple projectors and a large convex lens or a concave mirror is old and famous. The conventional methods, however, require diffusers to show continuous motion parallax, which decays the quality of the image. To solve this problem we use a convex lens array with no gaps between the lenses, which realizes continuous motion parallax without diffusers. The convex lens array, however, has to produce images without aberration to show the viewer stable 3D images. It is hard and expensive to realize such lens arrays without gaps between the component lenses. To produce images with little aberration in a simple format, the author proposes the optical system where each component lens makes the parallel light rays instead of creating an image by keeping the distance between the LCD surface and the lens array the same as the focal distance of the component lenses. To create an image, we use a large convex-type Fresnel lens, which has been used only for the purpose of distributing multiview images to each viewpoint in the conventional multi-projection systems. Fresnel lens, receiving parallel light from the lens array, creates a floating real image at its focal distance and attains distribution of multiview images at the same time. With this configuration we can create images with little aberration even when we use a lens array composed of simple convex-type Fresnel lenses widely available with low prices.

Range imaging technology: new developments and applications for people identification and tracking

Abstract: Range Imaging (RIM) is a new suitable choice for measurement and m odeling in many different applications. RIM is a fusion of tw o different technologies. According to the terminology, it integrates distance measurement as we ll as imaging aspects. The distance measurement principle is dominated by the time-of-flight principl e while the imaging array (e.g. CMOS sensor) enables each pixe l to store also the distance towards the corresponding object point. Du e to the technology’s relativel y new appearance on the market , with a few different realizations, the knowledge of its capabilities is very low. In this paper we present our investigations on the range imagi ng camera SwissRanger TM (realized by the Swiss Center for Electronics and Microtechnology - CSEM). Different calibration procedures are performed, including a photogrammetric camera calibration and a distance system calibration with respect to the reflectivity and the distance itself. Furthermore we report about measurement applications in the field of surveillance and biometrics. In particular, range imaging data of moving people are analyzed, to i dentify humans, detect their movements and recover 3D trajectories.

Perspex machine VIII: axioms of transreal arithmetic

Abstract: Transreal arithmetic is a total arithmetic that contains real arithmetic, but which has no arithmetical exceptions. It allows the specification of the Universal Perspex Machine which unifies geometry with the Turing Machine. Here we axiomatise the algebraic structure of transreal arithmetic so that it provides a total arithmetic on any appropriate set of numbers. This opens up the possibility of specifying a version of floating-point arithmetic that does not have any arithmetical exceptions and in which every number is a first-class citizen. We find that literal numbers in the axioms are distinct. In other words, the axiomatisation does not require special axioms to force non-triviality. It follows that transreal arithmetic must be defined on a set of numbers that contains{-∞,-1,0,1,∞,p} as a proper subset. We note that the axioms have been shown to be consistent by machine proof.

Two-dimensional statistical linear discriminant analysis for real-time robust vehicle-type recognition

Abstract: Automatic vehicle Make and Model Recognition (MMR) systems provide useful performance enhancements to vehicle recognitions systems that are solely based on Automatic License Plate Recognition (ALPR) systems. Several car MMR systems have been proposed in literature. However these approaches are based on feature detection algorithms that can perform sub-optimally under adverse lighting and/or occlusion conditions. In this paper we propose a real time, appearance based, car MMR approach using Two Dimensional Linear Discriminant Analysis that is capable of addressing this limitation. We provide experimental results to analyse the proposed algorithm's robustness under varying illumination and occlusions conditions. We have shown that the best performance with the proposed 2D-LDA based car MMR approach is obtained when the eigenvectors of lower significance are ignored. For the given database of 200 car images of 25 different make-model classifications, a best accuracy of 91% was obtained with the 2D-LDA approach. We use a direct Principle Component Analysis (PCA) based approach as a benchmark to compare and contrast the performance of the proposed 2D-LDA approach to car MMR. We conclude that in general the 2D-LDA based algorithm supersedes the performance of the PCA based approach.

Autostereoscopic display technology for mobile 3DTV applications

Abstract: Mobile TV is now a commercial reality, and an opportunity exists for the first mass market 3DTV products based on cell phone platforms with switchable 2D/3D autostereoscopic displays. Compared to conventional cell phones, TV phones need to operate for extended periods of time with the display running at full brightness, so the efficiency of the 3D optical system is key. The desire for increased viewing freedom to provide greater viewing comfort can be met by increasing the number of views presented. A four view lenticular display will have a brightness five times greater than the equivalent parallax barrier display. Therefore, lenticular displays are very strong candidates for cell phone 3DTV. Selection of Polarisation Activated MicrolensTM architectures for LCD, OLED and reflective display applications is described. The technology delivers significant advantages especially for high pixel density panels and optimises device ruggedness while maintaining display brightness. A significant manufacturing breakthrough is described, enabling switchable microlenses to be fabricated using a simple coating process, which is also readily scalable to large TV panels. The 3D image performance of candidate 3DTV panels will also be compared using autostereoscopic display optical output simulations.

Identification of in-field defect development in digital image sensors

Abstract: Although solid-state image sensors are known to develop defects in the field, little information is available about the nature, quantity or development rate of these defects. We report on and algorithm and calibration tests, which confirmed the existence of significant quantities of in-field defects in 4 out of 5 high-end digital cameras. Standard hot pixels were identified in all 4 cameras. Stuck hot pixels, which have not been described previously, were identified in 2 cameras. Previously, hot-pixels were thought to have no impact at short exposure durations, but the large offset of stuck hot pixels will degrade almost any image and cannot be ignored. Fully-stuck and abnormal sensitivity defects were not found. Spatial investigation found no clustering. We tracked hot pixel growth over the lifetime of one camera, using only normal photographs. We show that defects develop continually over the lifetime of the sensor, starting within several months of first use, and do not heal over time. Our success in tracing the history of each defect confirms the feasibility of using automatic defect identification to analyze defect response and growth characteristics in a multitude of cameras already in the field, without performing additional experiments or requiring physical access to the cameras.

High-speed line-scan camera with digital time delay integration

Abstract: Dealing with high-speed image acquisition and processing systems, the speed of operation is often limited by the amount of available light, due to short exposure times. Therefore, high-speed applications often use line-scan cameras, based on charge-coupled device (CCD) sensors with time delayed integration (TDI). Synchronous shift and accumulation of photoelectric charges on the CCD chip - according to the objects’ movement - result in a longer effective exposure time without introducing additional motion blur. This paper presents a high-speed color line-scan camera based on a commercial complementary metal oxide semiconductor (CMOS) area image sensor with a Bayer filter matrix and a field programmable gate array (FPGA). The camera implements a digital equivalent to the TDI effect exploited with CCD cameras. The proposed design benefits from the high frame rates of CMOS sensors and from the possibility of arbitrarily addressing the rows of the sensor’s pixel array. For the digital TDI just a small number of rows are read out from the area sensor which are then shifted and accumulated according to the movement of the inspected objects. This paper gives a detailed description of the digital TDI algorithm implemented on the FPGA. Relevant aspects for the practical application are discussed and key features of the camera are listed.

Veiling glare: the dynamic range limit of HDR images

Abstract: High Dynamic Range (HDR) images are superior to conventional images However, veiling glare is a physical limit to HDR image acquisition and display We performed camera calibration experiments using a single test target with 40 luminance patches covering a luminance range of 18,619:1 Veiling glare is a scene-dependent physical limit of the camera and the lens Multiple exposures cannot accurately reconstruct scene luminances beyond the veiling glare limit Human observer experiments, using the same targets, showed that image-dependent intraocular scatter changes identical display luminances into different retinal luminances Vision's contrast mechanism further distorts any correlation of scene luminance and appearance There must be reasons, other than accurate luminance, that explains the improvement in HDR images The multiple exposure technique significantly improves digital quantization The improved quantization allows displays to present better spatial information to humans When human vision looks at high-dynamic range displays, it processes them using spatial comparisons

Crosstalk in automultiscopic 3-D displays: Blessing in disguise?

Abstract: Most of 3-D displays suffer from interocular crosstalk, i.e., the perception of an unintended view in addition to intended one. The resulting "ghosting" at high-contrast object boundaries is objectionable and interferes with depth perception. In automultiscopic (no glasses, multiview) displays using microlenses or parallax barrier, the effect is compounded since several unintended views may be perceived at once. However, we recently discovered that crosstalk in automultiscopic displays can be also beneficial. Since spatial multiplexing of views in order to prepare a composite image for automultiscopic viewing involves sub-sampling, prior anti-alias filtering is required. To date, anti-alias filter design has ignored the presence of crosstalk in automultiscopic displays. In this paper, we propose a simple multiplexing model that takes crosstalk into account. Using this model we derive a mathematical expression for the spectrum of single view with crosstalk, and we show that it leads to reduced spectral aliasing compared to crosstalk-free case. We then propose a new criterion for the characterization of ideal anti-alias pre-filter. In the experimental part, we describe a simple method to measure optical crosstalk between views using digital camera. We use the measured crosstalk parameters to find the ideal frequency response of anti-alias filter and we design practical digital filters approximating this response. Having applied the designed filters to a number of multiview images prior to multiplexing, we conclude that, due to their increased bandwidth, the filters lead to visibly sharper 3-D images without increasing aliasing artifacts.

Free Viewpoint Image Generation Using Multi-Pass Dynamic Programming

Abstract: Ray-Space is categorized by Image-Based Rendering (IBR), thus generated views have photo-realistic quality. While this method has the performance of high quality imaging, this needs a lot of images or cameras. The reason why that is Ray-Space requires various direction's and position's views instead of 3D depth information. In this paper, we reduce that flood of information using view-centered ray interpolation. View-centered interpolation means estimating view dependent depth value (or disparity map) at generating view-point and interpolating that of pixel values using multi-view images and depth information. The combination of depth estimation and interpolation realizes the rendering photo-realistic images effectively. Unfortunately, however, if depth estimation is week or mistake, a lot of artifacts appear in creating images. Thus powerful depth estimation method is required. When we render the free viewpoint images video, we perform the depth estimation at every frame. Thus we want to keep a lid on computing cost. Our depth estimation method is based on dynamic programming (DP). This method optimizes and solves depth images at the weak matching area with high-speed performance. But scan-line noises become appeared because of the limit of DP. So, we perform the DP multi-direction pass and sum-up the result of multi-passed DPs. Our method fulfills the low computation cost and high depth estimation performance.

Gait analysis in forensic medicine

Abstract: We have combined the basic human ability to recognize other individuals with functional anatomical and biomechanical knowledge, in order to analyze the gait of perpetrators as recorded on surveillance video. The perpetrators are then compared with similar analyses of suspects. At present we give a statement to the police as to whether the perpetrator has a characteristic gait pattern compared to normal gait, and if a suspect has a comparable gait pattern. We have found agreements such as: limping, varus instability in the knee at heel strike, larger lateral flexion of the spinal column to one side than the other, inverted ankle during stance, pronounced sagittal head-movements, and marked head-shoulder posture. Based on these characteristic features, we state whether suspect and perpetrator could have the same identity but it is not possible to positively identify the perpetrator. Nevertheless, we have been involved in several cases where the court has found that this type of gait analysis, especially combined with photogrammetry, was a valuable tool. The primary requisites are surveillance cameras recording with sufficient frequency, ideally about 15 Hz, which are positioned in frontal and preferably also in profile view.

The normalized compression distance and image distinguishability

Abstract: We use an information-theoretic distortion measure called the Normalized Compression Distance (NCD), first proposed by M. Li et al., to determine whether two rectangular gray-scale images are visually distinguishable to a human observer. Image distinguishability is a fundamental constraint on operations carried out by all players in an image watermarking system. The NCD between two binary strings is defined in terms of compressed sizes of the two strings and of their concatenation; it is designed to be an effective approximation of the noncomputable but universal Kolmogorov distance between two strings. We compare the effectiveness of different types of compression algorithms? in predicting image distinguishability when they are used to compute the NCD between a sample of images and their watermarked counterparts. Our experiment shows that, as predicted by Li's theory, the NCD is largely independent of the underlying compression algorithm. However, in some cases the NCD fails as a predictor of image distinguishability, since it is designed to measure the more general notion of similarity. We propose and study a modified version of the NCD to model the latter, which requires that not only the change be small but also in some sense random with respect to the original image.