This PDF file contains the editorial “Organizing Your Information” for OE Vol. 48 Issue 03

We recorded a hologram in a photorefractive LiNbO₃:Fe crystal with a two-color recording technique in transmission geometry. The holographic recording involves a light with wavelength 633 nm for interference, and a light with wavelength 532 nm for exciting. The short wavelength light excites more charges so that the holographic recording is affected, and the storage capacity (M-number) and the sensitivity vary accordingly. We found that the optimized intensity ratio of the interference lights and the exciting light is between 50 and 60. In the optimized conditions, the M number and the sensitivity are enhanced by 43 and 35%, respectively. Because the crystal we used is a typical iron-doped LiNbO₃ crystal and the short wavelength exciting light source is inexpensive now, the proposed method is easy to be achieved in most holographic systems.

A novel photonic technique of 60-GHz millimeter-wave subcarrier generation base on four-wave mixing effect in a semiconductor optical amplifier (SOA) and a dual-frequency Brillouin fiber laser configuration is proposed. In this system, two new harmonic components with six times spacing of the microwave source frequency are created when an optical signal, generated by carrier-suppressed intensity modulation, is launched into the SOA. The two residual modulation sidebands are then suppressed by stimulated Brillouin scattering process, and the leaved idlers provide an millimeter-wave subcarrier signal.

Fabrication and characterization of an integrated 3-D linear taper are reported, which is based on (111) silicon-on-insulator for efficient coupling between a single-mode fiber and planar photonic devices. The fabrication process involved wafer bonding, anisotropic etching, and dry etching. The 3-D taper measured 75 μm in length. The input waveguide, 5.6×5 μm² in dimensions of the end facet, was compressed to dimensions of 0.6×1.8 μm² of the end facet in the output waveguide. The measured average net transmission loss was 0.52 dB at the wavelength of 1.55 μm.

We propose a fast subpixel motion estimation algorithm for the H.264 advanced video coding (AVC) standard. The algorithm utilizes the correlation of the spatial interpolation effect on the full-pixel motion estimation best matches between different block sizes in order to reduce the computational cost of the overall motion-estimation process. Experimental results show that the proposed algorithm significantly reduces the CPU cycles in the various motion estimation schemes by up to 16% with similar rate-distortion performance when weighed up against the H.264/AVC standard.

Nonparametric statistical modeling of background and foreground has been widely used for moving foreground segmentation from video sequences. In this work, a simple metric is presented to evaluate the performance of various foreground models. The proposed metric allows us to test the robustness of the foreground model to the motion and deformation of the moving foreground. Experiments are performed on five typical foreground models, showing that the proposed metric is effective.

We propose a new intra prediction algorithm that can improve the coding efficiency of the H.264/AVC standard. In the proposed algorithm, the prediction signal for each macroblock (MB) is formed by using all the previously reconstructed pixels of the current frame. Experimental results show that the proposed algorithm can reduce the bit rate by 1.36% to 9.07%, as compared with the conventional intra prediction, while the average PSNR is not decreased.

A frame grouping and merging scheme of sub-sequences to estimate precise camera motion parameters over long omnidirectional sequences is addressed. In this triplet-based approach, the back-projection errors of corresponding points by the camera model for grouping subsequences are used. This process accounts for the relative reliability of the estimated camera model in field of view. Experimental results show that the proposed method solves the problem inherent in sequential algorithm: its susceptibility to drift over long sequences.

A method to characterize the vibrating shape of a bimorph deformable mirror by measurement using a laser displacement sensor is presented. The displacement of the deformable mirror surface at a point is recorded as a frequency response function. The surface profile is synthesized from the response functions obtained at multiple points on a deformable mirror by postprocessing. The method has a wide frequency range and high-frequency resolution, suitable to investigate the resonance behavior of a deformable mirror. Although the spatial resolution is low compared with an interferometer, it is sufficient to identify low-spatial-order shapes at resonance frequencies. The accuracy of the displacement measurement is estimated to be comparable with interferometric data.

Many geodetic and industrial instruments utilize precise angle determination. Metrological calibration and testing of such devices is very complicated, due to the large number of angular position values generated, and often cannot be accomplished using classical means of angle determination. Therefore a special comparator (test rig) capable of calibration geodetic and other angle-measuring instruments and implementing multiple references for angle determination has been created. Several modernized optical devices were used on the test rig, together with the implementation of a special calibration method. In this paper the construction of the rig, modernization of the optical devices, calibration, and some results of measurements are presented.

Commercial glossmeters are usually provided with a gloss reference. In some cases the reference is derived from black glass. Light reflection from the reference depends both on the complex refractive index and on the surface quality, such as the roughness and the cleanness of the reflecting surface, of the black glass. If we want to obtain accurate data on gloss or the minute gloss mottling of an object, we must be certain of the optical properties and surface quality of the reference in regard to reliable gloss analysis. In this study the complex refractive index of a gloss reference, as a function of wavelength, was measured with an ellipsometer, and the reflectance at standardized gloss measurement angles was calculated for the case of nonpolarized light. We show that the gloss reference has a substantial extinction coefficient, which must be considered in accurately describing the reflectance of the reference. Furthermore, the gloss variation of the reference was detected with a diffractive-optical-element-based glossmeter by scanning the reference surface at normal light incidence. We observed that the reference had rather high gloss mottling and consider the reasons for it.

Measuring objects with a high variation range of surface reflectivity is challenging for any optical method: This paper addresses a high dynamic range scanning technique that can measure this type of object. It takes advantage of one merit of a phase-shifting algorithm: pixel-by-pixel phase retrieval. For each measurement, a sequence of fringe images with different exposures are taken: the brightest ones have good fringe quality in the darkest areas while the darkest ones have good fringe quality in the brightest areas. They are arranged from brighter to darker (i.e., from higher exposure to lower exposure). The final fringe images, used for phase retrieval, are produced pixel-by-pixel by choosing the brightest but unsaturated corresponding pixel from one exposure. A phase-shifting algorithm is employed to compute the phase, which can be further converted to coordinates. Our experiments demonstrate that the proposed technique can successfully measure objects with high dynamic range of surface reflectivity variation.

This paper examines the possibility of constructing deformable mirrors for adaptive optics with a large number of degrees of freedom from silicon wafers with bimorph piezoelectric actuation. The mirror may be used on its own, or as a segment of a larger mirror. The typical size of one segment is 100 to 200 mm; the production process relies on silicon wafers and thick film piezoelectric material deposition technology; it is able to lead to an actuation pitch of the order of 5 mm, and the manufacturing costs appear to grow only slowly with the number of degrees of freedom in the adaptive optics.

Image rendering of spatial light modulators (SLMs) is often degraded by various effects. Some efficient methods to get around nonuniformity, nonlinearity, and remanence and improve image rendering are presented. Optical laboratory results are presented for an analog ferroelectric liquid-crystal SLM using a high-speed camera. Focus is made on a preprocessing compensation method, using a spatial-dependant correction table. A high-speed "on-the-fly" implementation is also suggested.

A noninterferometric configuration for an optical phase-shifter module (OPSM) is presented and demonstrated, which is a key component in phase-shifted optical quantization (PSOQ) systems. In a PSOQ system employing such an OPSM, the input electrical analog signal is applied on two LiNbO₃ intensity modulators in parallel, and the OPSM takes the outputs of the intensity modulators as its input and yields N-channel optical outputs, which are thresholded to generate digitized values of the input analog signal. The feasibility of this OPSM configuration is demonstrated by a proof-of-principle PSOQ experiment, in which a 2.5-GHz single tone is applied to the modulators and 16 transmission curves are recorded. Based on these transmission curves, software sampling indicates that an effective number of bits equal to 4.17 is attainable for a frequency as high as 2.5 GHz. Benefits of such an OPSM are easier control and high precision of desired phase shifts.

The primary objective of the Decihertz Interferometer Gravitational Wave Observatory (DECIGO) mission is to detect and observe black holes' and galactic binaries' gravitational waves (GWs), at frequencies ranging from 10^-2 to 10¹ Hz (from 0.1 to 100 s in averaging time). This low-frequency range is inaccessible to ground-based interferometers, due to unshieldable background noise and to the fact that ground-based interferometers are limited to a few kilometers in length. Our research is focused on efforts to stabilize semiconductor, Nd:YAG, and fiber lasers, for use as GW detectors' optical sources. In present-day and future detectors, frequency and phase noise may place certain limitations on sensitivity and stability. Our goal is primarily to design robust experiments. In this report, we compare existing methods: Faraday, Faraday peak, and saturated absorption spectroscopy. In these, the laser frequency is stabilized to Rb as an atomic frequency reference by a feedback-loop control system. From the frequency stability of these models, we can predict the characteristics of the three systems through dynamic stability analysis, by analyzing the dynamic Allan variance. We find the optical frequency stability, expressed as the Allan deviation (the square root of the Allan variance), to be 3.3×10^-11, 2.9×10^-12, and 1.2×^-12 in the respective methods.

A pH sensor is developed based on the long-period grating (LPG) inscribed into a single-mode–multimode–single-mode (SMS) structure in order to measure the pH in seawater. The LPG is formed by using a focused CO₂ laser at LPG's critical period (1 mm). The transmission characteristics are similar to that of a multimode fiber. However, the SMS structure has a higher sensitivity because mode coupling only happens between the fundamental mode and higher-order modes in the SMS structure. The pH-sensitive hydrogel is made by the thermal cross-link of poly vinyl alcohol and poly acrylic acid. This sensor has been utilized in seawater pH sensing in the range of 6–8. Experiments have confirmed that the sensor is sensitive and repeatable.

The fiber Bragg grating has been widely used in sensors. We have studied the spectral properties of uniform Bragg gratings in photonic crystal fibers under transverse pressure. By the finite element method, the relation between the birefringence and the pressure was simulated in bare photonic crystal fibers. The results show that the birefringence is changed even under weak pressure. A new method based on polarization-dependent loss is presented for measuring pressure. The maximal amplitude of the polarization-dependent loss varies linearly in the pressure range of 0 to 4 MPa with a slope of 0.75 MPa^-1.

A novel structure of an optical fiber pressure sensor based on a fiber Bragg grating (FBG) and metal bellows is presented. Due to the novel structure, the sensitivity is improved to 48 pm/kPa. The relationship between Bragg wavelength and the applied pressure is derived. Experimental data indicates that there is a good linear relation between the Bragg wavelength shift and the applied pressure. This sensor can be utilized in low-pressure measurement.

Optical resonant filters represent an important alternative to achieve optical notch filtering. Enhancing the filtering properties by new designs for these filters is an essential aspect because they can be a key component in important systems, such as optical communications, bio-photonics applications, and optical sensors. The proposed designs depend on a multilayer waveguide that can function as a confining medium and an antireflection coating as well, which are not achieved simultaneously by a single-layer waveguide. Zeroing the Fresnel reflection, which is the background reflection in the reflection spectra of the optical resonant filter, is highly desirable for better filtering properties.

Switchable dual-wavelength fiber lasers with orthogonal polarizations, based on the semiconductor optical amplifier at room temperature, are proposed. One polarization-maintaining fiber Bragg grating (PMFBG) is used as the wavelength-selective filter in three different laser configurations. Owing to the polarization-dependent loss of the PMFBG, the laser can be designed to operate in stable dual-wavelength or wavelength-switching modes with a wavelength spacing of 0.336 nm at room temperature by adjusting a polarization controller in each laser configuration. The amplitude variation over nearly half an hour is less than 0.1 dB for both wavelengths in each laser configuration, which is more stable than that of erbium-doped fiber lasers with similar configuration.

An optical packet compression for data samples with different bit patterns and scaling degrees is proposed and demonstrated. The compression stages are composed of optical modulators, 3-dB couplers and fiber delay lines. Erbium-doped fiber amplifiers are applied only between pairs of compression stages. An experiment of 16-bit packet size with compression from 150 MHz to 2.5 GHz is also reported. The proposed scheme achieved low costs in terms of optical components while preserving the signal pattern and scaling degree integrity. It is also shown that the scaling monitoring is an important performance issue for optical communications systems.

We provide an analysis for the performance of a multiwavelength optical code-division multiple-access (MW-OCDMA) network when the system is working above the nominal transmission rate limit imposed by passive encoding-decoding operation. We address the problem of overlapping in such a system and how it can directly affect the bit error rate (BER). A unified mathematical framework is presented under the assumption of one-coincidence sequences with nonrepeating wavelengths. A closed form expression of the multiple access interference limited BER is provided as a function of different system parameters. Results show that the performance of the MW-OCDMA system can be critically affected when working above the nominal limit, an event that can happen when the network operates at a high transmission rate. In addition, the impact of the derived error probability on the performance of two newly proposed medium access control (MAC) protocols, the S-ALOHA and the R³T, is also investigated. It is shown that for low transmission rates, the S-ALOHA is better than the R³T, while the R³T is better at very high transmission rates. In general, it is postulated that the R3T protocol suffers a higher delay mainly because of the presence of additional modes.

The author proposes the design of a novel switchless variable optical delay line (VFODL) for radio frequency (rf) signals that combines an analog mode VFODL with a discrete-state VFODL. This hybrid VFODL proposes the use of a laser with smart, fast (e.g., 1 ns) wavelength tuning in combination with flat-top passband design spatially and temporally dispersive optical elements to simultaneously provide both long and short time delays, thus finely covering a wide nanoseconds delay range. An example design indicates a 9.6-ns time delay range with 6-ps increments, giving 1600 independent delays using a no-moving-parts signal-processing structure. VFODL applications include rf antenna array processing, radar testing, and precision electrical timing systems.

It is well established that optical devices and components are more advantageous than their electronic counterparts because of inherent parallelism in optics. Basically electronics are found to be very unsuitable in high speed (above gigahertz) data processing systems whereas tremendous operational speed (in the range of terahertz) can be achieved with the help of optics. The parallelism of optics and the properties of low loss transmission make optics a powerful technology for digital computing and processing and in long-range communications. Again it is well established that logic gates are the basic building blocks of any computing or data processing system. Therefore, any optical data processor needs suitable optically run logic gates. A method of conducting phase-modulated all-optical logic gates is proposed. Here we will exploit the advantages of phase modulation not only in processing but also in encoding as well decoding also.

A technique for 3-D selective imaging of sound sources is described analytically and demonstrated experimentally. One-dimensional recordings of the acoustic field is measured using laser vibrometry. By applying digital holographic and tomographic algorithms to the acquired 1-D data, the full 3-D complex amplitude is reconstructed. The use of multiple frequencies in the spectral content of the acoustic field gives a number of advantages: higher spatial resolution, less noise in the reconstructed image, less sensitivity to noise in the measurements, and the possibility to perform selective imaging. Theory for all three steps—the measurement of sound using light, numerical propagation of waves, and finally the tomographic reconstruction in the process are given. In the experiment, the positions of three ultrasound sources are accurately determined and two different types of transducers are distinguished from each other. This multiwavelength technique could show to be a useful addition to optoacoustic imaging.

Two new types of phenanthrenequinone-doped poly(methyl-methacrylate) (PQ/PMMA) are proposed and fabricated for volume holographic recording. In these materials, the matrix of PQ/PMMA is replaced by the copolymers, which are composed of either poly(trimethylolpropane triacrylate-co-methyl methacrylate) or poly(2-phenoxyethyl acrylate-co-methyl methacrylate), respectively. With the chemical analyses of the response of these materials with respect to light exposure, the physical mechanism of the holographic recording in those copolymer samples is investigated. In addition, the holographic characteristics, including the dynamic range and sensitivity, have been measured. The experimental results demonstrate that the modification of the monomer components can enhance holographic characteristics of PQ/PMMA.

A performance metric Δ is described that can be used to determine the parameters of a partially coherent beam for near-optimal performance of a free-space optical link. The metric is defined as the mean received irradiance minus the standard deviation of the irradiance, and maximizing Δ balances the effects of beam spread and scintillation. An analytic form of Δ is developed for a Gaussian Schell-model beam, where the optimization parameter is the transverse coherence length. Comparisons of the metric performance are made with the more conventional probability-of-fade metric. The metric Δ is applied to determine the characteristics of the optimized coherence length as a function of a variety of link parameters and scenarios. In general, the optimized coherence length tends to decrease with increasing turbulence strength and propagation distance but increases with wavelength, although the behavior of specific scenarios can vary.

We propose a procedure that can improve the coding efficiency of bilevel image compression schemes that are based on chain coding techniques and entropy coders. The two chain coding techniques analyzed in this study employ three and eight symbols, respectively, to code the outline of object shapes in bilevel images. Once we have the chain code representation of a contour, we group the original chain symbols in blocks of two or more symbols to build new extended alphabets from which we can obtain more efficient compression rates by using a statistical model that best represents our new alphabets and entropy coding. The coding efficiency of every new alphabet is measured by its entropy, average length in bits per original source symbol, and the total length in bits that represent the compressed image. We compare the compression efficiency between our two chain coding techniques using our three entropy coding parameters for symbol groupings of different sizes. In our experiments, we also show how our blocking approach produces an entropy value that is lower than its expected theoretical counterpart for a given order of symbol groupings and, as a consequence, the average and total lengths are expected to follow a decreasing pattern as well. Finally, we compare our best encoding model against the Joint Bilevel Image Experts Group's JBIG and JBIG2 bilevel image compression standards.

Objective video quality measurement has become an important issue, as multimedia services are now widely available over the Internet and other wireless communication media. Traditionally, professional CRT monitors have been used to measure subjective video quality. However, the majority of users have LCD, plasma display panel (PDP), or consumer-graded CRT monitors. We compared the subjective video quality of various TV and LCD PC monitors. Subjective tests were performed with a wide range of video sequences using different monitors, and their correlations were analyzed. Although there were high correlations among the various display monitors, care should be taken in selecting a monitor for certain applications.

A novel algorithm for 3-D tomographic reconstruction is proposed. The proposed algorithm is based on multiresolution techniques for local inversion of the 3-D Radon transform in confined subvolumes within the entire object space. Directional wavelet functions of the form Ψ^j_m,n(x)=2^j/2Ψ(2jw_m,nx) are employed in a sequel of double filtering and 2-D backprojection operations performed on vertical and horizontal reconstruction planes using the method suggested by Marr and others. The densities of the 3-D object are found initially as backprojections of coarse wavelet functions of this form at directions on vertical and horizontal planes that intersect the object. As the algorithm evolves, finer planar wavelets intersecting a subvolume of medical interest within the original object may be used to reconstruct its details by double backprojection steps on vertical and horizontal planes in a similar fashion. Reduction in the complexity of the reconstruction algorithm is achieved due to the good localization properties of planar wavelets that render the details of the projections with small errors. Experimental results that illustrate multiresolution reconstruction at four successive levels of resolution are given for wavelets belonging to the Daubechies family.

Linear space-invariant image restoration algorithms often introduce ringing effects near sharp intensity transitions. A ringing metric to evaluate the quality of images restored using iterative deconvolution algorithms is presented. According to the types of ringing artifacts, two ringing metrics are used to assess the restored images based on the analysis of ringing artifacts. An overall ringing metric is presented based on the two ringing metrics. The experimental results validate that the proposed method performs well over a wide range of restoration image ringing level assessments. Consequently, the proposed model has a good agreement with the ratings from an observer in subjective experiments.

We propose a method to detect and localize story-related subject captions in news video. Most caption detection and localization algorithms attempt to detect as many captions as possible; however, a news picture may contain many types of captions that are unrelated to the story. To facilitate fast and accurate access to news video content, a method for detecting and localizing the story-related caption is necessary. This paper addresses two problems in texture-based caption detection and localization: the time-consuming computation of features, and the clutter of caption detection results. We address these problems by first identifying the subject caption region based on the frequency of text occurrence. Then, we detect the subject caption frame as it first appears onscreen. Finally, the texture-based caption localization procedure is performed on the subject caption region in subject captions' beginning frames. Using this method decreases the computation time significantly. Additionally, the unrelated types of text are also filtered out, and only the subject caption is detected and localized. Experimental results show that the proposed method can quickly and robustly detect subject captions from news video.

A novel method using a multiple-symbol lookup table (mLUT) is proposed to fast-skip the ac coefficients (codewords) not needed to construct a dc image from MPEG-1/2 video streams, resulting in fast thumbnail generation. For MPEG-1/2 video streams, thumbnail generation schemes usually extract dc images directly in a compressed domain where a dc image is constructed using a dc coefficient and a few ac coefficients from among the discrete cosine transform (DCT) coefficients. However, it is required that all codewords for DCT coefficients should be fully decoded whether they are needed or not in generating a dc image, since the bit length of a codeword coded with variable-length coding (VLC) cannot be determined until the previous VLC codeword has been decoded. Thus, a method using a mLUT designed for fast-skipping unnecessary DCT coefficients to construct a dc image is proposed, resulting in a significantly reduced number of table lookups (LUT count) for variable-length decoding of codewords. Experimental results show that the proposed method significantly improves the performance by reducing the LUT count by 50%.

Live-wire-based image segmentation allows the user to select piece-wise globally optimal contour segments corresponding to a desired object boundary. However, the boundary extracted by live wire tends to deviate from the object boundary if other boundaries with strong gradient features are situated near a desirable object boundary. We present a new live-wire method based on the concept of embedded edge confidence to improve segmentation performance. Experimental results on medical images demonstrate the effectiveness of the proposed method.

A new multiple optical calibration system (MOCS)-based effective multiview video processing (MVP) is proposed. The proposed MOCS can be classified as intracamera processing and intercamera processing (IP). In IP, unequal brightness intensity is rectified by using a brightness difference map, and geometric distortion of the optical lens is corrected. In IP, several important factors, such as unequal camera sensor location, tilt degree of lens mount, and focal length, among multiple cameras are calibrated. Furthermore, an effective disparity matching algorithm, called an as adaptive-based disparity matching algorithm (A-DMA), is proposed for MVP. The proposed A-DMA enhances the performance of MVP in terms of accuracy of disparity estimation and computational overhead.

A pattern-recognition and encoding system has been developed for a biochip platform using shaped hydrogel sensors batch produced via photolithography. Each sensor shape is fashioned with a unique pattern of dots that makes it identifiable to a pattern recognition system. By linking the sensor's function to its shape, "random" arrays can be created (i.e., arrays that do not require sensors to be located at specific positions). Random arraying can be quickly and cost-effectively achieved via self-assembly methods. Pattern-recognition software was written to perform automated recognition of micrographs exhibiting fluorescing sensors. As a test of the recognition process, an array of shape-encoded DNA sensors was fabricated using lithography. Fluorescent micrographs were taken of a DNA-sensing experiment, and then processed with the pattern-recognition software. The results show that this process is quite viable with 98% recognition accuracy of the nondefective sensors in both images.

Although iris recognition is one of the most accurate biometric technologies, it has not yet been widely used in practical applications. This is mainly due to user inconvenience during the image acquisition phase. Specifically, users try to adjust their eye position within small capture volume at a close distance from the system. To overcome these problems, we propose a novel iris image acquisition system that provides users with unconstrained environments: a large operating range, enabling movement from standing posture, and capturing good-quality iris images in an acceptable time. The proposed system has the following three contributions compared with previous works: (1) the capture volume is significantly increased by using a pan-tilt-zoom (PTZ) camera guided by a light stripe projection, (2) the iris location in the large capture volume is found fast due to 1-D vertical face searching from the user's horizontal position obtained by the light stripe projection, and (3) zooming and focusing on the user's irises at a distance are accurate and fast using the estimated 3-D position of a face by the light stripe projection and the PTZ camera. Experimental results show that the proposed system can capture good-quality iris images in 2.479 s on average at a distance of 1.5 to 3 m, while allowing a limited amount of movement by the user.

A new method for 3-D object recognition is proposed. The method combines structured light coding with the joint fractional transform correlator (JFrTC).When projecting sinusoidal grating fringes onto the surface of 3-D objects, the fringes are distorted due to the height of the objects. Then the distorted fringe patterns are regarded as the joint input image used for correlation recognition. Simultaneously, to achieve JFrTC, we can replace either (or both) Fourier transforms in the joint transform correlator (JTC) by a fractional Fourier transform. Four architectures of JFrTC with the distorted fringe pattern and four JFrTC with the intensity map are analyzed in detail. By comparing these different structures, a JFrTC with distorted fringe patterns, especially for the FrT-FT architecture, is suggested to improve the discernment performance, to more decrease the widths of the correlation peaks, and to flexibly change the intensity and position of correlation peaks. Both mathematical analysis and computer simulation results are presented in support of the proposed idea.

The performance of face recognition systems has much been burdened by head pose variation. To solve this problem, 3-D face recognition systems that make use of multiple views and depth information have been suggested. However, without an accurate head pose estimation, the performance improvement of 3-D face recognition systems under pose variations remains limited. Previous research on 3-D head pose estimation has been conducted in 3-D space, where the estimation complexity is high. Also it is difficult to incorporate those salient 2-D face features for effective estimation. We propose a novel iterative 3-D head pose estimation method incorporating both 2-D and 3-D face information. To verify the effectiveness, we apply the proposed method to 3-D face modeling and recognition systems with adaptation to various 3-D face data acquisition devices. Our experimental results show that the proposed method can be very effective in terms of modeling and recognition applications, particularly on combining different kinds of acquisition devices, which use different coordinates of origin and scale.

Vision-based inspection systems have been widely investigated for the detection and classification of defects in various industrial product. We present a new defect detection algorithm for scale-covered steel billet surfaces. Because of the availability of various kinds of steel, presence of scales, and manufacturing conditions, the features of billet surface images are not uniform. In particular, scales severely change the properties of defect-free surfaces. Moreover, the various kinds of possible defects make their detection difficult. In order to resolve these problems and to improve the detection performance, two methods are proposed. First, undecimated wavelet transform and vertical projection profile are presented. Second, a method for detecting the variations in the block features along the vertical direction is proposed. The former method can effectively detect vertical line defects, and the latter can efficiently detect the remaining defects, except the vertical line defects. The experimental results conducted on billet surface images obtained from actual steel production lines show that the proposed algorithm is effective for defect detection of scale-covered steel billet surfaces.

This paper describes application-oriented text localization and character segmentation algorithms in images. The target text in our application includes many unclear characters due to poor environment as well as the fact that their positions are variable in the images. Consequently, it is difficult to expect a high success rate when using existing text localization algorithms that have been developed for generic texts. Therefore, it is necessary to develop a new text localization algorithm. We propose (1) a coarse algorithm for detecting top and bottom boundaries, (2) a fitness function that is used to decide the true text among the text candidates, (3) two kinds of presegmentation algorithms for calculating the fitness function, and (4) a blank-detecting algorithm that determines whether the text is upside down or not. By the proposed algorithms, input upside-down text is rotated automatically without using any supervised or unsupervised learning methods; further, character segmentation can be done in the process of selecting the true text. To evaluate the algorithms, image data captured by the installed recognition system at Pohang Steel Company (POSCO) are used, and experimental results show that the proposed algorithms are fast and reliable.