The term thermal printing includes a very broad technological basis of printing an image with the use of thermal energy. The printing process involves one of the following mechanisms: (1) transfer of dyes, (2) transfer of pigments or dyes with a carrier, or (3) formation of dyes on a substrate. Starting in the late fifties thermal printing has been used in the textile business. This technology was widely used later in printing tickets, receipts and bar codes and in fax machines. The application of this printing technology in color hard copies was pioneered by Dai Nippon and first commercialized by Sony in the Mavica electronic photographic system in the late eighties. Since then many companies have participated, in various capacities, in the thermal printing business. At this time there is no doubt that thermal dye transfer printing with the use of resistive or laser printhead gives the highest photographic image quality in digital printing as compared to an optical printing system with silver halide materials.

Liquid immersion development of electrostatic images involves the electrophoretic motion of charged toners and counter-ions in a spatially non-uniform and time-varying electric field originating from the imagewise charge on the photoreceptor. This paper summarizes the recent coordinated theoretical and experimental investigation of electrophoresis under a condition closely simulating the actual development process. The roles of chase density and mobility in the development process are elucidated by the theory of space-charge-limited transport. Their values are determined by a novel characterization technique, and are shown to be consistent with the prediction from the electrical double-layer model.

We describe general requirements of organic photoreceptors used for electrophotographic applications. A typical organic photoreceptor consists of a charge generation layer (CGL) and a charge transport layer (CTL). The separation of the generation and transport functions has permitted the optimization of both layers in terms of sensitivity as well as device lifetime. Dual-layer and negative-charged/hole- transporting organic photoreceptors are of primary interest in device applications, which are the focus here. The spectral sensitivity, charge mobilities, and the general characteristics of charge acceptance and photoinduced discharge of these materials are discussed. In flash exposure of electrostatic latent image, reciprocity failures reduce the sensitivity of photoreceptors and are key issues associated with digital electrophotography. We discuss reciprocity failures relating to Langevin recombination and perhaps exciton-exciton annihilation at high exposure conditions. Other related issues such as latent image spread and fatigue are also discussed.

In recent years, ink jet has emerged as one of the mainstream printing technologies. Since its market inception in 1985, Hewlett-Packard's thermal ink jet technology (TIJ) has evolved progressively from a 12 nozzle 96 dpi print head to a 300 nozzle 600 dpi print head. TIJ has made rapid progress enabling it to print text output on plain paper that challenges laser printers, and realistic photographic images that rival silver halide, at a low consumer price. Thermal ink jet technology continues to enjoy a greater unit market share than any other digital printing technology and all other ink jet technologies combined. The driving forces for the advancement of TIJ have been better, faster, and cheaper printers for consumers. These goals involve key attributes such as ink performance (gamut, sharpness, fastness), minimum deliverable colorant (drop volume), rate of colorant delivery (firing frequency, nozzle integration, firing chamber volume), and print engine cost per unit throughput. In this paper, key technology challenges for TIJ will be outlined. New materials and new processes that are required for the advancement of thermal ink jet printing are discussed. Recently, competing ink jet methods have (re-)emerged, notably piezoelectric ink jet. References will be made to piezoelectric ink jet when appropriate.

We present a new method to fabricate silicon microfluidic channels by through-hole etching with subsequent planarization. The method is based on etching out the deep grooves through a perforated silicon carbide membrane, followed by sealing the membrane with plasma-enhanced chemical vapor deposition (PECVD). Low-pressure-chemical-vapor- deposited (LPCVD) polysilicon was used as a sacrificial layer to define the channel structure and only one etching step is required. This permits the realization of planarization after a very deep etching step in silicon and offers the possibility for film deposition, resist spinning and film patterning across deep grooves. The process technology was demonstrated on the fabrication of a monolithic silicon microchannel structure for thermal inkjet printing. The Ta-Al heater arrays are integrated on the top of each microchannel, which connect to a common on-chip front-end ink reservoir. The fabrication of this device requires six masks and no active nozzle-to-chip alignment. Moreover, the present micromachining process is compatible with the addition of on-chip circuitry for multiplexing the heater control signals. Heat transfer efficiency to the ink is enhanced by the high thermal conductivity of the silicon carbide in the channel ceiling, while the bulk silicon maintains high interchannel isolation. The fabricated inkjet devices show the droplet sizes of 20 - 50 micrometer in diameter with various channel dimensions and stable ejection of ink droplets more than 1 million.

The passivation layers in the thermal inkjet printhead are not only crucial for the life time, but also for thermal characteristics of the thermal printhead device. In this paper, therefore, the experimental and simulated results are presented to investigate the effects of the passivation layers on the bubble onset characteristics and temperature distribution. The experiments were carried out by use of the open pool test to measure the bubble onset voltage. The effects of the different passivation layer structures and the loop resistance on the bubble onset voltage are studied. Moreover, a one-dimensional heat transfer model is employed in this paper to understand the temperature distribution across the device structure.

In a thermal bubble inkjet printhead, thin film resistor plays the most important role. The thin film resistor must fire about 1 billion times during the device operation. During firing, temperatures are up to 400 degrees Celsius and the duration is on the order of microseconds. In this situation, the resistance stability and mechanical property of the thin film resistor are critical for fabrication. In this study, resistivity and stress behavior related to a wild range of Ta- Al compositions are examined.

Electronic imaging systems based on solid-state image sensors have received tremendous attention in the past few years. The many advantages of the solid-state image sensors have opened up a wide variety of new imaging possibilities. Applications, ranging from high resolution, spaceborne electronic camera to PC peripheral camera, were costly, it not impossible, using the earlier technology. This paper will describe the technologies of charge-coupled device (CCD) and CMOS image sensors. A review of the recent advancements and industry trends will be given. Challenges in sensor performance, device scaling, and fabrication processes will also be discussed. The paper will conclude with a description of some applications and opportunities for the solid-state image sensors.

The spectral sensitivity of a CCD camera is a prominent electro-optical characteristic for color image formation. In this paper, spectral estimation of color CCD cameras generally involves the assessment of overall spectral sensitivities, which are composed of those of optics, filters, and CCD sensors. To this end, we develop an automatic testing/measurement system with an optical platform. And, the spectral characterization of a CCD camera imaging process is presented in the approach of vector representation for the estimation problem. This technique examines the modulated incident colored lights, which are regarded as a set of color stimuli candidates, to be the effective color stimuli, such that they can be successfully applied to generic estimation algorithms. To show the feasibility of this work, a realization of a basic set of color stimuli candidates is generated in the numerical and graphical forms. In our approach the use of a spectrophotometer is needed only for the system calibration. It would be pleasing that this technique will practically simplify the complexity of the spectral estimation problem of CCD cameras.

Commercial electronic still picture cameras need a low-cost, systematic method for evaluating the performance. In this paper, we present a measurement method to evaluating the dynamic range and sensitivity by constructing the opto- electronic conversion function (OECF), the fixed pattern noise by the peak S/N ratio (PSNR) and the image shading function (ISF), and the spatial resolution by the modulation transfer function (MTF). The evaluation results of individual color components and the luminance signal from a PC camera using SONY interlaced CCD array as the image sensor are then presented.

Nowadays, 3D data are popularly performed in computer, and 3D browsers manipulate 3D model in the virtual world. Yet, till now, 3D digitizer is still a high-cost product and not a familiar equipment. In order to meet the requirement of 3D fancy world, in this paper, the concept of a low-cost 3D digitizer system is proposed to catch 3D range data from objects. The specified optical design of the 3D extraction is effective to depress the size, and the processing software of the system is compatible with PC to promote its portable capability. Both features contribute a low-cost system in PC environment in contrast to a large system bundled in an expensive workstation platform. In the structure of 3D extraction, laser beam and CCD camera are adopted to construct a 3D sensor. Instead of 2 CCD cameras for capturing laser lines twice before, a 2-in-1 system is proposed to merge 2 images in one CCD which still retains the information of two fields of views to inhibit occlusion problems. Besides, optical paths of two camera views are reflected by mirror in order that the volume of the system can be minified with one rotary axis only. It makes a portable system be more possible to work. Combined with the processing software executable in PC windows system, the proposed system not only saves hardware cost but also processing time of software. The system performance achieves 0.05 mm accuracy. It shows that a low- cost system is more possible to be high-performance.

A speckle image correlation method is proposed for the application to 3D topography at ultra-high speeds. With random densely-seeded laser speckles projected onto an object to be measured, taking snapshots from two different viewing angles using a pair of CCD cameras, we are able to correlate the two speckle images in real time, 30 Hz, by a sparse-array compressed-image correlation algorithm. As soon as the correlation is established, a fast triangulation reconstruction gives the 3D coordinates of the speckles on the object in the global coordinate system, and thus the topography of the surface of the object is obtained. This method has an accuracy of plus or minus 0.2 pixels and a resolution equal to that of its cameras. Most image correlation algorithms suffer from their lack of speed. The method discussed here has overcome this barrier and made wide applications possible. Real-time correlation enables applications in areas such as on-line inspection, non-still object measurement, and instant 3D model acquisition.

Conventional interpolation techniques, including pixel replication, bilinear interpolation and spline based methods have been popularly used in commercial applications. It is often desired that the interpolation process is able to accurately and fastly increase the spatial resolution of an image at an arbitrary aspect ratio and with sharp edges. These do not generally happen to conventional algorithms, which will consume a great effort to interpolate an image at an arbitrary ratio and tend to blur edges or introduce blocking artifacts. In this correspondence, we propose an interpolation system that is more general than the existing ones. The method is based on the fact that the Hartley transform (HT) at an arbitrary frequency can be expressed as a weighted sum of its Discrete Hartley transform (DHT) coefficients. These weights can be suitably approximated so that the HT is very nearly the sum of (1) a few dominant terms of the sum of the DHT coefficients, and (2) the DHT of a new sequence obtained by multiplying the original sequence with a saw-tooth function. If we take the inverse discrete Hartley transform (IDHT) of an image; then by using the algorithm described above, the spatial sample at an arbitrary location can be fastly computed by the fast Hartley transform (FHT) algorithm. In addition to retaining the computational efficiency of the FHT algorithm, experimental results have revealed that the proposed system preserves the sharp edge of the original image.

This paper describes the design of a flexible, pipelined general image processor (GIP) using VHDL to model the top level design and functional blocks consisting of histogram [1,2,3,4,5,6], modification, convolution, halftone, error diffusion, and threshold. GIP was simulated to have a processing speed of 70 Mpixels/second. A four pixel wide image data path is used so a clock of 17.5 MHz can be used. Mentor Graphics tool suites were used to perform the simulation and synthesis of the design. The total number of gates in 1.2 (mu) CMOSN gate array was estimated to be 236 K gates, less than 1 million transistors.

In this paper, we present a Bayesian decision-based neural networks (BDNN) for handwritten Chinese character recognition. The proposed Self-growing Probabilistic Decision-based Neural Networks (SPDNN) adopts a hierarchical network structure with nonlinear basis functions and a competitive credit-assignment scheme. Our prototype system demonstrates a successful utilization of SPDNN to the handwriting of Chinese character recognition on the public databases, CCL/HCCR1 and in house database (NCTU/NNL). Regarding the performance, experiments on three different databases all demonstrated high recognition (86 - 94%) accuracy as well as low rejection/acceptance (6.7%) rates. As to the processing speed, the whole recognition process (including image preprocessing, feature extraction, and recognition) consumes approximately 0.27 second/character on a Pentium-100 based personal computer, without using hardware accelerator or co-processor.

Two major issues in image coding are the effective incorporation of human visual system (HVS) properties and the effective objective measure for evaluating image quality (OQM). In this paper, we treat the two issues in an integrated fashion. We build a JND model based on the measurements of the JND (Just Noticeable Difference) property of HVS. We found that JND does not only depend on the background intensity but also a function of both spatial frequency and patten direction. Wavelet transform, due to its excellent simultaneous Time (space)/frequency resolution, is the best choice to apply the JND model. We mathematically derive an OQM called JND_PSNR that is based on the JND property and wavelet decomposed subbands. JND_PSNR is more consistent with human perception and is recommended as an alternative to the PSNR or SNR. With the JND_PSNR in mind, we proceed to propose a wavelet and JND based codec called JZW. JZW quantizes coefficients in each subband with proper step size according to the subband's importance to human perception. Many characteristics of JZW are discussed, its performance evaluated and compared with other famous algorithms such as EZW, SPIHT and TCCVQ. Our algorithm has 1 - 1.5 dB gain over SPIHT even when we use simple Huffman coding rather than the more efficient adaptive arithmetic coding.

This paper extends and modifies Classified Vector Quantization (CVQ) to solve the problem of inverse halftoning. The proposed process consists of two phases: the encoding phase and decoding phase. The encoding procedure needs a codebook for the encoder, while the decoding process requires another codebook for the decoder. The difference between an input vector and its corresponding codeword is included to reconstruct a gray-scale image. The experiments show that our algorithm is robust to the filter which is used to generate an error-diffused image. Compared with other available techniques, our approach has the better image quality. The main contribution of this paper is that it opens another area of application for VQ.

This paper presents one part of our work on 'hierarchical visualization.' Here, we propose an efficient hierarchical structure for the visualization of 3D volumetric data, where for instance, data compression techniques are embedded into the hierarchical structure. The hierarchical structure consists of two layers. The first hierarchy roughly visualizes the whole lossy compressed data so that we can quickly understand and interpret the outline of the whole 3D volumetric dataset. This step allows users to choose the desired parts of the 3D volumetric data. The second hierarchy is used for scrutinizing the detail of the chosen parts using lossless compressed data. To implement the hierarchical structure we propose a partitioning algorithm for 3D volumetric data based on principal component analysis. With principal component analysis, the original 3D volumetric space can be divided into a number of 3D volumetric blocks such that each block contains similar data. Analyzing these specific blocks and taking advantage of an octree hierarchy, from the whole 3D volumetric dataset we can easily access only the parts which are required for display. The proposed partitioning algorithm should be very useful in reducing the amount of rendering, which means that we will need only to precisely render the blocks required for display, and not the whole 3D volumetric dataset.

Since it is impossible to automatically parse a general video, we investigated an integrated solution for the content-based news video indexing and the retrieval. Thus, a specific structural video such as news video is parsed, because it is included both temporal and spatial characteristics that the news event with an anchor-person is iteratively appeared, a news icon and a caption are involved in some frame, respectively. To extract automatically the key frames by using the structured knowledge of news, the model used in this paper is consisted of the news event segmentation, caption recognition and search browser module. The following are three main modules represented in this paper: (1) The news event segmentation module (NESM) for both the recognition and the division of an anchor-person shot. (2) The caption recognition module (CRM) for the detection of the caption-frames in a news event, the extraction of their caption region in the frame by using split-merge method, and the recognition of the region as a text with OCR software. 3) The search browser module (SBM) for the display of the list of news events and news captions, which are included in selected news event. However, the SBM can be caused various searching mechanisms.

The cut detection in a simple video sequence is relatively simple. But it becomes very difficult and complex when there are the video sequences which has special effects, i.e. gradual transition scene in the sequential frames or scenes with flash between the sequential frames. Twin comparison method were proposed earlier, but the methods also have the problem of the double calculation for applying threshold. This research review the method of detecting cuts known thus far, then analyzes and evaluates performance and reliability of those methods by experiments. This research proposes a new algorithm of detecting cuts using motion vector, and compares it with other existing methods. In addition, we simulate the synthesized video sequences that include special effects in order to prove an effectiveness of the proposed algorithm, and the results show good status.

The advent of media processors is facilitating the rapid implementation of multimedia processing algorithms in various applications. Multimedia processing presents challenges from the perspectives of both hardware and software. Media processing solutions include Intel MMX, embedded memory-based processor arrays, and special purpose media processors such as Trimedia. In general, there are two approaches for multimedia processor design namely: Dedicated and Programmable. The objective of this paper is to present a detailed categorization of the various architectural alternatives for multimedia processing based on the programmability of hardware and data distribution.

A method for obtaining intelligent behavior and influencing the shapes of artificial creatures following an evolutionary model is described. The creatures obtain intelligent behavior by interacting with the variable conditions in the environment where they live. Our algorithm proposes a way for the creatures to shirk from enemies, overcome obstacles, search for a mate and look after their own needs by using the five senses. The evolutional model used in our research is based on Genetic Algorithms (GA) and offers a novel way to generate new shapes from intelligent behaviors. This paper proposes a method to generate intelligent behaviors for and evolve the shape of artificial creatures by expanding our previously proposed evolutionary model.

This paper investigates the feeding of receiver (recording paper) by capstan roller for thermal transfer type color printing. In thermal transfer color printing systems, this kind of feeding mechanism including a motor driven capstan roller and a pinch roller is commonly employed. It repeatedly advances receiver back and forth to overlay different color image separations. Since this mechanism handles the receiver by friction, accurate receiver feed registration without slippage is important factor for high printing quality. The surface characteristic of capstan roller can influence the receiver feed precision and the color registration. So the experimental apparatus with different kinds of capstan surface is established to examine this effect upon receiver feed registration. It can find the method of avoiding mis- registration. The results presented will be very helpful for mechanism designers to make better choice in developing this kind of receiver feed mechanism.

This paper describes an on-line inspection system especially dedicated for the OPC drum production line. This inspection system can be divided into three portions. The first one is the automatic robot that enables the OPC drums to enter the inspection area successively and gives an indication of the test result to the operator. The second is the image capture system that is comprised of an optical module and an A/D converter board. The A/D converter board digitizes the image and transfers the image data for further processing. The third portion is the personal computer having a dedicated algorithm that processes the image data and transmits the test result to the robot immediately. The robot also can be operated either by automatic or manual operations and has some safety designs to prevent the operator from being injured. Due to the complexity of coating surface, the inspection algorithm has been developed to adapt for various flaws of the OPC drum. Except for a few of the low contrast spots, most of the defects can be identified effectively by this inspection system. This inspection system is available not only for OPC drum but has potential for the other industrial applications as well.

An effective and flexible data conversion controller for ink jet printing is proposed in this paper. It provides the capability of processing firing data to meet the requirements of both firing sequence and firing timing of the cartridge. This print data is arranged according to the nozzle layout of a selected print cartridge and the printing direction. Then output the arranged data to the firing driver of print cartridge. A hardware architecture is proposed to serve the needs of data conversion. It is an effective design that consists of two appointed selections of nozzle firing sequence and fixed nozzle pad mapping. Moreover, the proposed architecture provides programmable firing timing and it can be switched between black or color printing mode. The core technology of this data conversion hardware is now applied to an ASIC of ink jet fax engine by OES/ITRI.

Based upon the purpose of prolonging the lifetime of an ink- jet head, we have attempted to apply double pulse to a heater during a drop ejection cycle. There are five parameters (pulse widths and voltages and the time delay) to define the double pulse. Taguchi's method is used to investigate the effects of these parameters. The experimental results show that under the optimized condition with (18 V, 2 microseconds), (14 V, 4 microseconds) and delay time 12 microseconds double pulse can prolong the life of ink jet heads up to an averaged drop counts of 4.8 X 10⁸ as compared with that of 3.0 X 10⁸ using single pulse. Also, improper parameters can dramatically shorten the life of an ink-jet head, or disturb the drop formation.

Three-dimensional technique was active in the past 40 years and a variety of systems had been developed. Among these systems, field-sequential stereoscopic TV system is the most economical and practical one. A flicker-free field-sequential stereoscopic TV system presented in this paper will provide you with a true vivid 3D effect and will be more practical compared with other systems. Frame memory design is the most important part in this flicker-free field-sequential stereoscopic TV system. I will give a minute description of frame memory design in this paper.

A new holographic technique has been used to make a compact, accurate and reliable POS scanner. The holo-window technology permits compact POS scanner optical scanning in horizontal plan while maintaining excellent performance in changing the scan direction, equalizing the scan velocity and collecting the signal light. The holo-window design and fabrication in the holographic optical element (HOE) for such a compact POS scanner are described in this paper. Additionally this new horizontal side scanning possesses large depth of field (greater than 10 inches), allows the grocery items to be scanned horizontally thus eliminating the commonly experienced carpal tunnel syndrome (CTS) hand injuries of the checkers. This newly designed POS scanner has been recognized by industry as the standard for the future POS scanning configuration.

Reproducing colors with rich saturation, from illuminating objects, is usually recognized as an essential issue for CCD camera imaging. In this paper, we propose a colorimetric calibration scheme regarding self-luminous images for CCD cameras. And, an efficient algorithm to generate highly saturated color stimuli is devised for investigating the CCD camera performance of image reproduction. In this scheme, a set of color samples containing highly saturated colors is generated, from an advanced CRT, as color stimuli for colorimetric characterization. To demonstrate the effectiveness of this algorithm, a realization of color samples, uniformly distributed in CIE LAB, are presented for illustration.

In this paper, an Atomic Force Microscope (AFM) with laser lever and its application inspecting microprofile of optical surface are presented. The nanometer-grade defects in optical and supersmooth surface can be tested by AFM. The microprofile drawings have the accuracy higher than (0.1 - 1 nm) in both vertical and lateral resolution. Some testing results of optical and supersmooth surfaces are shown in the paper.

The fractal dimension of image is an important quantitative character, which can be widely used in image analysis. But general algorithm according to the definition of Box Counting dimension needs more computation and cannot be real-time computed. In order to compute the fractal dimension of an image in real-time, an algorithm is designed in this paper by using morphological cover, according to the definition of Minkowski-Bouligand dimension. The result of experimental computation manifests that the fractal dimension of the image can be precisely and robustly computed by using the algorithm proposed in this paper. The architecture of real-time parallel implementation is also given.

With the immense size of images, compression has become a common way of minimizing the amount of storage necessary for images. This is also beneficial for transmission purposes. The Joint Photographic Experts Group (JPEG) standard is frequently used for still images. This standard is very flexible and many of the same algorithms can be used for video applications. Video applications require large amounts of data to be processed every second. Therefore, the following describes the hardware design of a chip allowing for high-speed compression. The design uses the JPEG algorithms and is targeted towards ASIC design. Further plans include use of field programmable gate arrays (FPGAs). The hardware design is based on grayscale images and only works with the raw image data.

A BGO array coupled to the PSPMT (Position Sensitive Photo- Multiplier Tube) is used as the image system to identify the position of the scintillation event. The newly developed PSPMT R5900-00-C8 (Hamamatsu Photonics, K. K) with metal channel dynodes and 8 cross plate anodes is the key component to determine the location and the energy of the scintillation event. The imaging instruments are arranged as follows: BGO array is coupled to the PSPMT. Each of the eight anode signals generated by the PSPMT (4 for X dimension and 4 for Y dimension) are amplified and offset-compensated. A trigger module provides a proper trigger and gating signal for the ADC module. A delay device is inserted between each output of the preamplifier and the input of the charge integrating ADC to adjust the timing delay introduced by the trigger module. The digitized signals captured by the ADC are recorded and processed by computer to locate the energy and position of the scintillation events. The imaging characteristics of the BGO- PSPMT combination have been evaluated. Results suggest that this imaging device may be very useful in applications requiring high spatial resolution radiation detection by a compact imaging device.

This paper investigates the question: What is the minimum number of colors required to represent color images in a computer monitor. We conduct experiment to perform JND partition along the 3-axis in L x y color space that colors in the same partition are indistinguishable to human perception. We also propose a color image quality measure based on the LMS cone perception sensitivity. The JND model is applied to design a fixed color palette and its performance is evaluated.