Real-time imaging

his paper presents a new motion estimation algorithm to improve the performance of the existing searching algorithms at a relatively low computational cost. We try to amend the incorrect and/or inaccurate estimate of motion with higher precision by using Kalman filter. We first obtain a measurement of motion vector of a block by using the existing searching scheme. We then generate the predicted motion vector utilizing the inter-block correlation in both spatial and temporal directions. In general, the motion correlation in spatial direction is different from that in temporal direction. Therefore, if we appropriately utilize the motion correlation in spatial direction and temporal direction, the better performance would be achieved. In this paper, we will propose an adaptive Kalman filter which utilizes a model switching mechanism to select correct motion model. Simulation results show that the proposed technique can efficiently improve the motion estimation performance.

Shipwreck SZD1 is the biggest of eleven wrecks resting in the Regalica River (West Pomerania, Poland), inventoried during the project "Underwater Ethnoarchaeology of the Lower Oder. Preliminary Research on Wrecks in Selected Sections of the River" (grant no. 2018/02/X/HS3/00475, National Science Centre, Poland). It has been an object of interest for explorers and scholars for years because previous investigations by means of sonar (stationary and side-scan) and interferometric echo-sounder had not allowed for identifying its type accurately and recognizing details visible on the recorded images. The authors of the paper reached the conclusion that it was necessary to investigate the wreck one more time by means of another kind of sonar – ARIS acoustic camera, which makes it possible to record underwater space in an analogous way to the optical video camera. The investigation was carried out in September 2019 on board of a research vessel "Echo 2", owned by ESCORT Sp. z o.o. During the survey, high-quality video records were made, making it possible not only to identify the type of the sunken vessel, but also to recognize its construction details. Moreover, the damage to the ship’s hull was observed, making it possible to enunciate hypotheses about why it sank. The sonar images also show many variously shaped objects resting on the river bed, both next to the wreck itself and at some distance from it. The authors assumed that at least some of these were the cargo of the sunken ship. In order to verify the hypotheses and suppositions, it would be necessary to carry out a diving reconnaissance on shipwreck SZD1 in the future.

his paper describes the application of digital image processing and pattern recognition techniques to assist in diagnosing neurological disorders. In medical practices the posture and movement of a human subject through his/her gait cycle contains information that is used by an experienced clinician to determine the mental health of a patient. An image processing based system can be used to automate this process and produce quanti®ed results. This is achieved by processing, extracting and classifying joint angle information from still images of a human subject's gait. Several new algorithms are devised to process and extract the required information from the images: a new edge extraction technique is used to assist in image segmentation; a technique based on correlating images from dierent parts of a gait is used to obtain complete information about a subject's posture; and an algorithm based on the Hough Transform is used to obtain the limb joint angles and swing distances of a subject. Joint angles and swing distances obtained from normal and patient subjects are then used in training and verifying classi®cations using a feed-forward neural network and a fuzzy clustering algorithm.

irect Fourier methods (DFM) of tomographic reconstruction have been investigated since the introduction of the technique in the early 1970s, although computerized tomograms Dempl oy the convoluted back projection method (CBPM) as the means of computing the reconstruction. The main difficulty with DFM is the required interpolation of radially sampled data to the Cartesian grid in order to allow a 2D Fourier transform to be performed on the frequency plane data, and is at least part of the reason for the preference for CBPM. However, CBPM is difficult to implement quickly on even highly specialized hardware. In contrast to this, the main computational components of the DFM are a series of 1D Fourier transforms on the projection data followed by a 2D Fourier transform subsequent to a radial-to-Cartesian coordinate conversion. The Fourier transforms can be computed very efficiently using a fast Fourier transform (FFT) algorithm implemented in digital signal processing hardware. This paper describes the development of a DFM reconstruction method. The issue of the frequency plane interpolation is addressed and the effect of the accuracy with which this is performed on image reconstruction quality is examined. It was found that good reconstruction quality could be obtained with a combination of zero-padding the projection data and simple bilinear interpolation of the frequency plane data array from radial to Cartesian coordinates. Hardware was developed based around the Sharp LH9124 FFT DSP chipset that is capable of computing a 2D 512 ϫ 512 resolution complex to complex Fourier transform in 32 ms. Thus, by implementing the DFM with this hardware it is possible to compute tomographic reconstructions in several tens of milliseconds, some two orders of magnitude faster than currently achieved with CBPM. This would be of value in a number of medical and industrial applications that require dynamic processes to be examined in real-time.

Atomic force microscopy (AFM) has become a powerful addition to the range of instruments available to probe the organization of lipid monolayers and bilayers. Currently, AFM is the only tool that can provide nanoscale topographic images of supported lipid membranes under physiological conditions, enabling researchers to resolve their detailed structure and to monitor their interaction with drugs, peptides and proteins. Here, we survey recent data obtained by our research groups that demonstrate the power of the technique for exploring peptide–membrane interactions, with an emphasis on microbial lipopeptides and on tilted peptides. Copyright © 2008 John Wiley & Sons, Ltd.

Three-dimensional human head modeling is useful in video-conferencing or other virtual reality applications. However, manual construction of 3D models using CAD tools is often expensive and time-consuming. Here we present a robust and e cient method for the construction of a 3D human head model from perspective images viewing from di erent angles. In our system, A generic head model is rst used, then three images of the head are r equired to adjust the deformable contours on the generic mod e l t o m a k e i t more closer to the target head. Our contributions are a s follows. Our system uses perspective images that are more r ealistic than orthographic projection approximation used in earlier works. Also for shaping and positioning face organs, we present a method for estimating the camera f o cal length and the 3D coordinates of facial landmarks when the camera transformation is known. We also provide an alternative for the 3D coordinates estimation using epipolar geometry when the extrinsic parameters are absent. Our experiments demonstrated that our approach produces good and realistic results.

n this paper, we present a VLSI architecture for real-time image processing in quality control industrial applications: automation of the visual inspection phase of mechanical parts treated by the IFl uorescent Magnetic Particle Inspection method for structural-defect detection. The VLSI architecture implements a highly constrained neural network tailored for this specific application: the multi-layer perceptron with strictly local connections. The learning of the weights is performed off line by using the adaptive simulated-annealing algorithm. The neural network has been trained on real plant data: recognition results of the training and classification tasks compare favorably with those obtained by expert human operators. The VLSI architecture receives as input the image (taken on-line on the plant) of a mechanical part and it will find out if at least one structural surface defect is present. The VLSI architecture was optimized, through a set of transformations on the high-level VHDL specifications of the neural network algorithm, to reach real-time operating conditions. Following the proposed approach and the designed architecture, we designed and successfully tested a custom VLSI chip for the real-time implementation of the recognition task.

Atomic force microscopy (AFM) has become a powerful addition to the range of instruments available to probe the organization of lipid monolayers and bilayers. Currently, AFM is the only tool that can provide nanoscale topographic images of supported lipid membranes under physiological conditions, enabling researchers to resolve their detailed structure and to monitor their interaction with drugs, peptides and proteins. Here, we survey recent data obtained by our research groups that demonstrate the power of the technique for exploring peptide–membrane interactions, with an emphasis on microbial lipopeptides and on tilted peptides. Copyright © 2008 John Wiley & Sons, Ltd.

Several algorithms are reviewed for computing various types of wavelet transforms: the Mallat algorithm, the "a trous" algorithm and their generalizations by Shensa. The goal is 1) to develop guidelines for implementing discrete and continuous wavelet transforms efficiently, 2) to compare the various algorithms obtained and give an idea of possible gains by providing operation counts. The computational structure of the algorithms rather than the mathematical relationship between transforms and algorithms, is focused upon. Most wavelet transform algorithms compute sampled coefficients of the continuous wavelet transform using the filter bank structure of the discrete wavelet transform. Although this general method is already efficient, it is shown that noticeable computational savings can be obtained by applying known fast convolution techniques (such as the FFT) in a suitable manner. The modified algorithms are termed "fast" because of their ability to reduce the computational complexity per computed coefficient from L to log L (within a small constant factor) for large filter lengths L. For short filters, we obtain smaller gains: "fast running FIR filtering" techniques allow one to achieve typically 30% save in computations. This is still of practical interest when heavy computation of wavelet transforms is required, and the resulting algorithms remain easy to implement.

In this paper we present parallel algorithms to solve the problem of image restoration when the Point Spread Function is Space Variant. The problem has a very high computational complexity and it is very hard to solve it on scalar computers. The algorithms are based on the decomposition of the image spatial domain and on the solution of both constrained and uncostrained restoration subproblems of size smaller than the original. The main results can be summarized as follows: (a) the quality of restorations do not depend on the number of subdomains; (b) the uncostrained restoration is scalable and e cient even with a large number of processors while the constrained restoration is e cient for subdomains of more than 50 50 pixels. The numerical tests have been executed on a Cray T3E with 128 processors.

The image segmentation is the basic step in the image processing involved in the processing of medical images. Image segmentation is an essential but critical component in low level vision image analysis and, pattern recognition. It is one of the most difficult and challenging tasks in image processing which determines the quality of the final result of the image analysis[1]. This paper presents an approach to segment the medical images obtained from the mammograms It extracts the texture from the given image. Various image segmentation algorithms are discussed.

Abstrucf-Video delivery from a server to a client across a network is an important component of many multimedia applications. While delivering a video stream across a resource constrained network, loss of frames may be unavoidable. Under such circumstances, it is desirable to find a servcr transmission schedule that can efficiently utilize the network resources while maximizing the perceived quality-of-service (QoS) at the client. To address this issue, in this paper we introduce the notion of selective ,frume discurd at the server and formulate the opfimul selective frume discitrd problem using a QoS-based cost function. Given network bandwidth and client buffer constraints, we develop an O ( N log N ) algorithm to find the minimum number of frames that must be discarded in order to meet these constraints. The correctness of the algorithm is also formally established. Since the computational complexity of the optimal algorithm for solving the optimal selective frame discard problem is prohibitively high in general, we also develop several efficient heuristic algorithms for selective frame discard. These algorithms are evaluated using JPEG video traces.

W e present a real-time foveation system for remote and distributed surveillance. The system performs face detection, tracking, selective encoding of the face and background, and efficient data transmission. A Java-based client-server architecture connects a radial network of camera servers to their central processing unit. Each camera server includes a detection and tracking module that signals the human presence within an observation area and provides the 2D face coordinates and its estimated scale to the video transmission module. The captured video data are then efficiently represented in log-polar coordinates, with the foveation point centered on the face, and sent to the connecting client modules for further processing. The current set-up of the system employs active cameras that track the detected person, by switching between smooth pursuit and saccadic movements, as a function of the target presence in the fovea region. High reconstruction resolution in the fovea region enables the successive application of recognition/verification modules on the transmitted video without sacrificing their performance. The system modules are well suited for implementation on the next generation of Java-based intelligent cameras.

perception sensor for a mobile robot based on a reconfigurable and flexible parallel architecture is developed. Low-level image processing is implemented by using A reprogrammable FPGA technology, and high-level image processing is performed by a dedicated reconfigurable parallel architecture based on digital signal processor TMS320C40. Hierarchical LINDA permits user-friendly parallel programming and real-time application. To show that our perception system can adapt to any application, we undertake the parallelization of a 3D scene-reconstruction algorithm using a geometric method, to meet the application response time (10 Hz).

The next section of the paper presents an overview of the existing methods followed by the adopted technique for the extraction of tree ring boundaries. The 3D reconstruction is given in section 4 followed by the obtained results and discussion in section 5 and finally we conclude our work.

A real-time implementation of an approximation of the support vector machine decision rule is proposed. This method is based on an improvement of a supervised classification method using hyperrectangles, which is useful for real-time image segmentation. The final decision combines the accuracy of the SVM learning algorithm and the speed of a hyperrectanglesbased method. We review the principles of the classification methods and we evaluate the hardware implementation cost of each method. We present the combination algorithm which consists of rejecting ambiguities in the learning set using SVM decision, before using the learning step of the hyperrectangles-based method. We present results obtained using Gaussian distribution and give an example of image segmentation from an industrial inspection problem. The results are evaluated regarding hardware cost as well as classification performances.

econstruction of given noisy data is an ill-posed problem and a computationally intensive task. Nonlinear regularization techniques are used to find a unique solution under certain constraints. In our contribution we present a parallel mixed-signal architecture which solves this nonlinear problem with in microseconds. By connecting all parallel cells in a circular manner it is possible to process noisy data vectors of infinite length. This is achieved by virtually shifting the nonlinear adaptive filter kernel over the noisy data vector. Additionally, we focus on the interaction between theory, discretization, numerical simulations, macro-modeling, and analog VLSI implementation for a theoretically well understood class of computer vision in an exemplary and paradigmatic way. A one-dimensional (1D) experimental chip has been fabricated using 0.8 mm CMOS technology. On-chip measurements are shown to agree with results from numerical simulations. Results from applying the 1D chip to nonlinear smoothing of twodimensional image data will also be given correspondence.

econstruction of given noisy data is an ill-posed problem and a computationally intensive task. Nonlinear regularization techniques are used to find a unique solution under certain constraints. In our contribution we present a parallel mixed-signal architecture which solves this nonlinear problem with in microseconds. By connecting all parallel cells in a circular manner it is possible to process noisy data vectors of infinite length. This is achieved by virtually shifting the nonlinear adaptive filter kernel over the noisy data vector. Additionally, we focus on the interaction between theory, discretization, numerical simulations, macro-modeling, and analog VLSI implementation for a theoretically well understood class of computer vision in an exemplary and paradigmatic way. A one-dimensional (1D) experimental chip has been fabricated using 0.8 mm CMOS technology. On-chip measurements are shown to agree with results from numerical simulations. Results from applying the 1D chip to nonlinear smoothing of twodimensional image data will also be given correspondence.

In this paper, we present a low-complexity 'reversible variable length code' (RVLC) decoding scheme for MPEG-4 video that recovers more blocks and sometimes more macroblocks (MBs) from error propagation region of corrupted video packets, as compared to the MPEG-4 scheme. The remaining blocks and MBs are concealed. Simulation studies have been carried out to show that the proposed scheme achieves better data recovery, both in terms of peak-signal-to-noise ratio (PSNR) and perceptual quality. In addition, we present more conditions for error detection than those suggested in MPEG-4, discuss properties of error propagation in corrupted video packets, and provide the RVLC codeword structure. The use of RVLC code structure will enable the decoding in both directions by using only one code table, instead of separate tables used for forward and backward decoding. Since the scheme is purely decoder based, the compliance with the standard is fully maintained. It is very simple for the decoder to keep track of the additional blocks/MBs recovered by the proposed scheme. Therefore, the data recovery scheme suggested in MPEG-4 can still be used as default scheme.

video rate FLIR image processing system is described. Four TM5320C44 digital signal processors (DSPs) are used to form a pipe line-type parallel architecture. The processing system Ai mproves image quality by performing motion compensated frame analysis. Thirty frame per second throughput is achieved at standard TV resolution.

Antibody-fluorophore conjugates have high potential for the specific fluorescence detection of target cancer cells in vitro and in vivo. However, the antibody-fluorophore conjugates described to date are inappropriate for real-time imaging of target cells because removal of unbound antibody is required to reduce background fluorescence before quantifiable analysis by microscopy. In addition, clinical applications of the conjugates have been limited by persistent background retention due to their long systemic circulation and nonspecific uptake. Here we report fast and real-time near-infrared fluorescence imaging of target cancer cells using an antigen-responsive molecular "on-off" sensor: the fluorescence of trastuzumab-ATTO680 conjugate is dark (i.e., turned off) in the extracellular region, while it becomes highly fluorescent (i.e., turned on) upon binding to the target antigen HER2 on cancer cell surface. This molecular switch enables fast and real-time imaging of targe...

Real-Time Tracking of Moving Objects with an Active Camera T his article is concerned with the design and implementation of a system for real-time monocular tracking of a moving object using the two degrees of freedom of a camera platform. Figure-ground segregation is based on motion without making any a priori assumptions about the object form. Using only the first spatiotemporal image derivatives, subtraction of the normal optical flow induced by camera motion yields the object image motion. Closed-loop control is achieved by combining a stationary Kalman estimator with an optimal Linear Quadratic Regulator. The implementation on a pipeline architecture enables a servo rate of 25 Hz. We study the effects of time-recursive filtering and fixed-point arithmetic in image processing and we test the performance of the control algorithm on controlled motion of objects.

Recently, some video-based eye-gaze detection methods used in eye-slaved support systems for the severely disabled have been studied. In these methods, infrared light was irradiated to an eye, two feature areas (the corneal reflection light and pupil) were detected in the image obtained from a video camera and then the eye-gaze direction was determined by the relative positions between the two. However, there were problems concerning stable pupil detection under various room light conditions. In this paper, methods for precisely detecting the two feature areas are consistently mentioned. First, a pupil detection technique using two light sources and the image difference method is proposed. Second, for users wearing eye glasses, a method for eliminating the images of the light sources reflected in the glass lens is proposed. The effectiveness of these proposed methods is demonstrated by using an imaging board. Finally, the feasibility of implementing hardware for the proposed methods in real time is discussed.

Driver inattention is one of the main causes of traffic accidents. Monitoring a driver to detect inattention is a complex problem that involves physiological and behavioral elements. Different approaches have been made, and among them Computer Vision has the potential of monitoring the person behind the wheel without interfering with his driving. In this paper I have developed a system that can monitor the alertness of drivers in order to prevent people from falling asleep at the wheel. The other main aim of this algorithm is to have efficient performance on low quality webcam and without the use of infrared light which is harmful for the human eye. Motor vehicle accidents cause injury and death, and this system will help to decrease the amount of crashes due to fatigued drivers. The proposed algorithm will work in three main stages. In first stage the face of the driver is detected and tracked. In the second stage the facial features are extracted for further processing. In last stage the most crucial parameter is monitored which is eye's status. In the last stage it is determined that whether the eyes are closed or open. On the basis of this result the warning is issued to the driver to take a break.

I n this paper, we investigate methods of reducing the computational complexity of the discrete cosine transform (DCT) in a software video encoder. The number of DCT calculations may be reduced by modeling the distribution of zero blocks. We demonstrate that the reduction in computational complexity is variable and depends on the statistics of the video sequence. We propose a new adaptive algorithm that can maintain a near-constant reduction in complexity. The proposed algorithm performs well at converging to a ''target'' computational complexity, at the expense of a small reduction in image quality. This algorithm provides a flexible mechanism for managing computational complexity in a video encoder.

In this paper, we present a new shape analysis approach using the well-known wavelet transform and exploring shape representation by landmarks. First, we describe the approach adopted to represent the landmarks data as parametric signals. Then, we show the relation of the derivatives of Gaussian wavelet transform applied to the signal-to-differential properties of the shape that it represents. We present experimental results using real data to show how it is possible to characterize shapes through multiscale and differential signal-processing techniques in order to relate morphological variables with phylogenetic signal, environmental factors and sexual dimorphism. The goal of this research is to develop an effective wavelet transform-based method to represent and classify multiple classes of shapes given by landmarks.

Real-Time Vision-Based System for Textile Fabric Inspection T his paper presents an automatic vision-based system for quality control of web textile fabrics. The general hardware and software platform developed to solve this problem is presented and a powerful algorithm for defect inspection is proposed. Based on the improved binary, textural and neural network algorithms the proposed method gives good results in the detection of many types of fabric defects under real industrial conditions, where the presence of many types of noise is an inevitable phenomenon. A high detection rate with good localization accuracy, low rate of false alarms, compatibility with standard inspection tools and low price are the main advantages of the proposed system as well as the overall inspection approach.

A memory power optimization and performance exploration methodology based on high-level (C language) code transformations that allows the system designer to explore various data memory power, data memory area and performance trade-offs early in the design process of embedded multimedia systems is introduced. This exploration strategy is introduced for both single and multiprocessor environments. The latter requires partitioning of the application. After employing software transformations, the experimental results, obtained using four well-known motion estimation kernels provide an insight on the performance and energy consumption trade-offs, comparing memory hierarchies for the ARM programmable core and prove the validity of the proposed approach.

his paper presents a new, fast geometrical shape recognition technique based on the properties of cellular automata (CA). The VLSI implementation of the architecture developed for this Tpur pose is also presented. The digitized binary image of the geometrical shape is loaded onto a 2D CA grid. This binary image is the initial global state of the CA. The CA evolves in time until a final stable global state is reached. The geometrical shapes are classified into four different categories, according to the symmetries of their final stable global state, and are then recognized. Eleven geometrical shapes have been recognised using the proposed technique. The die size dimensions of the chip for a 8 ϫ 8 pixel image are 2.56 mm ϫ 2.70 mm = 6.91 mm 2 , and its maximum frequency of operation is 35 MHz. Targeted applications include classification and inspection tasks in industry.

Eye gaze can be a rich source of information to identify particular interests of human users. Eye gaze tracking has been largely used in different research areas in the last years, as for example in psychology, visual system design and to leverage the user interaction with computer systems. In this paper, we present an IR-based gaze tracking framework that can be easily coupled to common user applications and allows for real-time gaze estimation. Compared to other gaze tracking systems, our system uses only affordable consumer-grade hardware and still achieves fair accuracy. To evaluate the usability of our gaze tracking system, we performed a user study with persons of different genders and ethnicities.

Computer-assisted vision plays an important role in our society, in various fields such as personal and goods safety, industrial production, telecommunications, robotics, etc. However, technical developments are still rare and slowed down by various factors linked to sensor cost, lack of system flexibility, difficulty of rapidly developing complex and robust applications, and lack of interaction among these systems themselves, or with their environment. This paper describes our proposal for a smart camera with real-time video processing capabilities. A CMOS sensor, processor and, reconfigurable unit associated in the same chip will allow scalability, flexibility, and high performance.

The CANDELA project aims at realizing a system for real-time image processing in traffic and surveillance applications. The system performs segmentation, labels the extracted blobs and tracks their movements in the scene. Performance evaluation of such a system is a major challenge since no standard methods exist and the criteria for evaluation are highly subjective. This paper proposes a performance evaluation approach for video content analysis (VCA) systems and identifies the involved research areas. For these areas we give an overview of the state-of-the-art in performance evaluation and introduce a classification into different semantic levels. The proposed evaluation approach compares the results of the VCA algorithm with a ground-truth (GT) counterpart, which contains the desired results. Both the VCA results and the ground truth comprise description files that are formatted in MPEG-7. The evaluation is required to provide an objective performance measure and a mean to choose between competitive methods. In addition, it enables algorithm developers to measure the progress of their work at the different levels in the design process. From these requirements and the state-of-the-art overview we conclude that standardization is highly desirable for which many research topics still need to be addressed.

T he deconvolution of images obtained by means of optical-sectioning widefield fluorescence microscopy, is a relevant problem in biological applications. Several methods have been proposed in the last few years, with different degrees of success, to improve the quality of the images, but the data complexity and the computational cost remain a limiting factor in this problem. We present in this paper an approach to perform the deconvolution of threedimensional data obtained by fluorescence microscopy (widefield) based on the Projection onto Convex Sets theory. Initially, a brief review of the Projection onto Convex Sets theory is presented. In the restoration algorithm, we combine some constraint sets to restore the out-offocus blur, to reduce the Poisson noise due to the acquisition process, to retrieve the missing frequencies due to the transfer function of the optical system, and to prevent the regularization errors. Some examples using a phantom and real cell images are presented. The method demonstrates good performance in terms of both visual results and cost-benefit analysis.

A Real-Time Image Noise Reduction System for FLIR Applications video rate FLIR image processing system is described. Four TM5320C44 digital signal processors (DSPs) are used to form a pipe line-type parallel architecture. The processing system Ai mproves image quality by performing motion compensated frame analysis. Thirty frame per second throughput is achieved at standard TV resolution.

A Real-Time Image Noise Reduction System for FLIR Applications video rate FLIR image processing system is described. Four TM5320C44 digital signal processors (DSPs) are used to form a pipe line-type parallel architecture. The processing system Ai mproves image quality by performing motion compensated frame analysis. Thirty frame per second throughput is achieved at standard TV resolution.

Advanced video surveillance applications require two successive steps: image analysis and content understanding. The first step analyses and extracts the characteristics of the video sequence. It defines the regions or the objects of interest according to their spatial/temporal properties. This analysis results in a segmentation of the video sequence that is interpreted by the content understanding step according to the specific scenario and surveillance requirements. This paper addresses the image analysis problem for a video surveillance system. We propose a statistical model-based change detection technique that defines the areas of interest in the image. Each area is analyzed separately by integrating spatial and temporal descriptors in a multi-feature clustering algorithm. The selective procedure we propose minimizes the computational load and significantly improves the results provided by the change detection technique. We test this method on both indoor and outdoor surveillance sequences. All the results show a correct segmentation of the scene. Moreover each object defined in the segmentation is described in terms of its spatial and temporal properties. These results can represent a valid input for a later content understanding procedure in several surveillance scenarios.

Negotiating the World of Make-believe:-The Aesthetic Compass R egardless of how far one stretches the definition of virtual reality (VR) (hopefully not so far as to negate its scientific and technological identity), the field is one of high expectations and of many yet unattained goals (technical, scientific, communicational). Some of its hardware (the headset, the glove) indeed became available at bearable prices (as promised by the early innovative entrepreneurs). Tools for building VR applications help in the design and implementation of new products deployed in hard-core science research or, at the other end of the spectrum, in wide-audience entertainment. Progress in the design and production of faster processors, better I/O (i.e., increased bandwidth and speed), and simple, more robust operating systems lead many to predict that the maturity of computer technology will be established in the high performance specific to the domain of virtual reality. However, an issue more significant to the ascertainment and viability of VR has yet to be approached: the crudeness, not to say ugliness, of its means of expression, in particular, of its visual component. The disparity between technological progress, culturally acknowledged formal qualities translated into VR components, and the ability to generate new expressive forms of distinction is undermining the field. The more one has the chance to experience new virtual realities, some extremely ingenious from a computational viewpoint, the more one fears that the medium is under the curse of submediocrity, at least in respect to its aesthetic condition. One of the theses of this article is that in the VR domain, as in any other form of human activity, the efficiency of the activity depends upon its underlying aesthetics. The degree of acceptance of new scientific and technological means, methods, and perspectives reflects expectations of human practice more than anything else. The aesthetic component, as difficult to formalize and encode as it is, affects the legitimacy of innovative endeavors, which VR definitely is, by ultimately affecting its efficiency. Hype and hope mix easily in reality, but even more in the realm of the virtual. The expectation of a strict definition of VR might seem an issue of academic interest only. But once we address the encompassing aspect of effectiveness in this field-effectiveness being its raison d'étre-we rapidly understand that specificity, i.e., precise scientific and epistemological identity, is reflected in design expectations and requirements for VR applications. These can be seen as instantiations of aesthetic exigencies. 01995 Academic Press Limited

In this paper we propose novel methods to evaluate the performance of object detection algorithms in video sequences. This procedure allows us to highlight characteristics (e.g., region splitting or merging) which are specific of the method being used. The proposed framework compares the output of the algorithm with the ground truth and measures the differences according to objective metrics. In this way it is possible to perform a fair comparison among different methods, evaluating their strengths and weaknesses and allowing the user to perform a reliable choice of the best method for a specific application. We apply this methodology to segmentation algorithms recently proposed and describe their performance. These methods were evaluated in order to assess how well they can detect moving regions in an outdoor scene in fixed-camera situations.

This paper presents a feature-based object tracking algorithm using optical flow under the non-prior training (NPT) active feature model (AFM) framework. The proposed tracking procedure can be divided into three steps: (i) localization of an object-of-interest, (ii) prediction and correction of the object's position by utilizing spatio-temporal information, and (iii) restoration of occlusion using NPT-AFM. The proposed algorithm can track both rigid and deformable objects, and is robust against the object's sudden motion because both a feature point and the corresponding motion direction are tracked at the same time. Tracking performance is not degraded even with complicated background because feature points inside an object are completely separated from background. Finally, the AFM enables stable tracking of occluded objects with maximum 60% occlusion. NPT-AFM, which is one of the major contributions of this paper, removes the off-line, preprocessing step for generating a priori training set. The training set used for model fitting can be updated at each frame to make more robust object's features under occluded situation. The proposed AFM can track deformable, partially occluded objects by using the greatly reduced number of feature points rather than taking entire shapes in the existing shape-based methods. The on-line updating of the training set and reducing the number of feature points can realize a real-time, robust tracking system. Experiments have been performed using several in-house video clips of a static camera including objects such as a robot moving on a floor and people walking both indoor and outdoor. In order to show the performance of the proposed tracking algorithm, some experiments have been performed under noisy and low-contrast environment. For more objective comparison, PETS 2001 and PETS 2002 datasets were also used.

Antibody-fluorophore conjugates have high potential for the specific fluorescence detection of target cancer cells in vitro and in vivo. However, the antibody-fluorophore conjugates described to date are inappropriate for real-time imaging of target cells because removal of unbound antibody is required to reduce background fluorescence before quantifiable analysis by microscopy. In addition, clinical applications of the conjugates have been limited by persistent background retention due to their long systemic circulation and nonspecific uptake. Here we report fast and real-time near-infrared fluorescence imaging of target cancer cells using an antigen-responsive molecular "on-off" sensor: the fluorescence of trastuzumab-ATTO680 conjugate is dark (i.e., turned off) in the extracellular region, while it becomes highly fluorescent (i.e., turned on) upon binding to the target antigen HER2 on cancer cell surface. This molecular switch enables fast and real-time imaging of targe...

W e examine color quantization of images using trellis-coded quantization (TCQ) in both the RGB domain and YUV domain. Together with a simple dithering scheme, an eightbit trellis-coded color quantizer reproduces images that are visually indistinguishable from the 24-bit originals. It can be viewed as a predictive trellis-coded color quantization scheme. We also study trellis-coded vector quantization (TCVQ) for color quantization. Our proposed TCQ-based color quantization schemes are universal in the sense that no training or look-up table is needed. The complexity of TCQ is linear with respect to image size, making trellis-coded color quantization suitable for color printing, real-time interactive graphics and window-based display environment.

Real-time Defect Inspection of Textured Surfaces T exture analysis plays an important role in the automated visual inspection of textured images to detect their defects. For this purpose, model-based and feature-based methods are implemented and tested for textile images in a laboratory environment. The methods are compared in terms of their success rates in determining the defects. The Markov Random Field model is applied on dierent DSP systems for real-time inspection.

Clinching technology allows to join metal sheets by using a cold press. The quality control of joint points, a type of joint buttons, is performed by experts observing the shape of the joint button on the basis of their experience. VISICON IST Research and Development project (partially funded by the European Commission) has realized algorithms and a distributed quality control system for assessing in real time the production of clinched galvanized metal boards. The solution is based on computer vision, software engineering, process modeling, and knowledge representation through object-oriented modeling. The image processing algorithm is based on the G Transform. The transform presents interesting properties and is computationally cheap. The VISICON solution has been validated by using a large set of data and statistical analysis for the detection of joint buttons and for their quality assessment. The paper reports a description of the distributed real-time architecture of the VISICON computer based quality control system, the main aspects of the computer vision processing for quality assessment and the results of the validation phase.

학위논문(박사)--서울대학교 대학원 :공과대학 기계공학부,2019. 8. Lee, Junghoon.Commercial high-resolution optical microscopes are essential for microscopy imaging; however, they are expensive and bulky, which limits their use in point-of-care devices, resource-limited areas, and real-time imaging of a sample in a large apparatus. In this study, we report a novel compact lightweight submicron-resolution reflected and inverted optical microscope at low cost. Our technique utilizes the proximity of the image sensor to a commercial microscope objective lens for compactness of the microscope. The use of an image sensor with a small pixel size helps to reduce the information loss, which provides high-resolution images. Moreover, our technique offers a freedom to tailor the design of microscope according to the required resolution, cost, and portability for specific applications, which makes it a suitable candidate for affordable point-of-care devices. Unlike conventional reflected microscope, which consists of d...

T his work presents an automated shape inspection system for two-dimensional plates with variable luminance. The system, installed in the rolling mill of a steel factory, captures linear images of plates at high temperatures (700-12001C) while they are moving on a roll path. The main objective of the system is to capture and characterize the shapes of the heads and tails of the plates, in order to optimize the parameters of the rolling mill, and so to minimize wastage. To date, most of the research work done on illumination has focused on obtaining optimal illumination for the inspected objects. This work, however, deals with the particular case of objects that radiate their own light. This system automatically adapts itself to different levels of luminance using a mechanism that calculates the proper exposure time to acquire each image. The mechanism combines a feedforward control with an adaptive feedback control loop for the exposure time.