This PDF file contains the editorial “Designing Optical Systems, Part II: Optimization” for OE Vol. 43 Issue 08

An optical model for simulating light extraction efficiency of a GaN-based LED chip is presented. We propose the introduction of a periodic sharpening structure on the interface between the sapphire and the n-GaN of a GaN-based LED to obtain as high as 78% light extraction efficiency.

This PDF file contains the editorial “Advances in Pattern Recognition Algorithms, Architectures, and Devices” for OE Vol. 43 Issue 08

The detection and discrimination of targets in infrared imagery has been a challenging problem due to the variability of target and clutter (background) signatures. We discuss the application of a novel quadratic filtering method using missile seeker infrared closing sequences. Image filtering techniques are well suited for target detection applications, since they avoid the disadvantages of typical pixel-based detection schemes (such as segmentation and edge extraction). Another advantage is that the throughput complexity of the filtering approach, in the detection process, also does not vary with scene content. The performance of the proposed approach is assessed on several datasets, and the results are compared with that of previous linear filtering techniques. Since we can obtain the signature of some of the clutter "in the field" or during operation, we examine the impact of updating the filters to adapt to the clutter.

We introduce a new concept in pattern recognition that we call heterocorrelation. Contrary to standard approaches, heterocorrelation allows correlation of different images solely by modifying the filter's intensity transmissivity in the areas of greatest phase mismatch relative to the phase of the stored template. This approach can convert a single object recognition filter to a multiple object classification filter. We develop three algorithms and successfully test each of them for three completely different input cases: one that uses geometric input images with very little common edge information, one that uses these same images encoded with high spatial frequency information, and one that uses synthetic aperture radar (SAR) images that have almost no edge information. The third algorithm provides shift-invariant heterocorrelation with equalized in-class correlation peaks, and eliminates the possibility of any higher-order side peaks for the added in-class objects. We also demonstrate how the use of a heterocorrelation filter can improve the performance of conventional multiplexed filters, such as synthetic discriminant filters, as well as increase their template storage.

Standard image compression algorithms may not perform well in compressing images for pattern recognition applications, since they aim at retaining image fidelity in terms of perceptual quality rather than preserving spectrally significant information for pattern recognition. New compression algorithms for pattern recognition are therefore investigated, which are based on the modification of the standard compression algorithms to simultaneously achieve higher compression ratio and improved pattern recognition performance. This is done by emphasizing middle and high frequencies and discarding low frequencies according to a new distortion measure for compression. The operations of denoising, edge enhancement, and compression can be integrated in the same encoding process in the proposed compression algorithms. Simulation results show the effectiveness of the proposed compression algorithms.

Object shape in an image can vary for many reasons. Therefore, a goal of object recognition research is to create algorithms that are able to accurately recognize objects with variability in shape. This paper suggests recognizing shape through an assessment of different measures. There are two primary approaches to measuring shape variations for recognition: measure, compare, and match; and compare, measure, and match. In the first approach, attributes of the object shape are measured and compared with the same measured attributes of the template shape. This paper focuses on the second approach, which first compares the object and template jointly and then creates a normalized measure for matching. This approach is called multiple joint comparative normalized measures (MJCNM). Confidence in the match is shown to be better against certain shape variabilities on using MJCNM than on using just one shape measure. In particular, the MJCNM approach here uses matched-filter, Procrustes, partial-directed Hausdorff, and percent-pixels-same measures. An experimental result is given that demonstrates the implementation and usefulness of that approach.

A real-time implementation of a joint transform correlator by using JPEG-compressed reference images is proposed in order to solve storage problems and improve the time response of automatic target recognition systems. The correlation performance is studied quantitatively by using two types of images with different spatial-frequency contents. The simulation results show that in comparison with the compressed high-spatial-frequency images, the joint transform correlator using the compressed low-spatial-frequency reference image offers better recognition performance in that it is robust to noise and contrast difference for a wide range of compression levels.

We present a technique for multiwavelet analysis using a phase-encoded joint transfer correlator (JTC). The optoelectronic JTC is presented and analyzed as an effective tool for ultrafast wavelet analysis. By multiplexing reference wavelets in phase, multiwavelet analysis is performed in a single correlation. The architecture of the proposed system and computer simulations are presented.

The use of a laser diode in a joint transform correlator (JTC) is proposed. Effects of temporal coherence, spatial coherence, astigmatism, and elliptical beams are analyzed. Laser diodes can be used in JTC, and also can be used for the correlation process in a Vander Lugt correlator and reading a hologram. Even a light-emitting diode (LED) may be used for the purposes mentioned, if its exit facet is sufficiently small. Two methods for correcting a laser diode beam using a cylindrical lens or a pair of cylindrical lenses are also reviewed. A corrected laser diode can be used in JTC to totally eliminate the effect of astigmatism.

A technique of optical pattern recognition that combines the performance of a wavelet transformation and a multilevel composite filter is proposed. The essential advantage of this technique is distortion invariance. We show by digital simulation that this technique can successfully identify and discriminate complex biometric images, such as fingerprints distorted digitally by various types of distortions: rotation (up to 10 deg clockwise and counterclockwise), shift (up to 10% of the image size), occlusion, scaling (changes up to 5%), and pinch and punch (changes up to 10%). The wavelet transform is used for extracting crucial ridge information from tested images and eliminating its redundancies. The multilevel composite filter is generated by a simulated annealing algorithm.©

An adaptive target detection algorithm for forward-looking infrared (FLIR) imagery is proposed, which is based on measuring differences between structural information within a target and its surrounding background. At each pixel in the image a dual window is opened, where the inner window (inner image vector) represents a possible target signature and the outer window (consisting of a number of outer image vectors) represents the surrounding scene. These image vectors are then preprocessed by two directional highpass filters to obtain the corresponding image gradient vectors. The target detection problem is formulated as a statistical hypotheses testing problem by mapping these image gradient vectors into two linear transformations, P₁ and P₂, via principal component analysis (PCA) and eigenspace separation transform (EST), respectively. The first transformation P₁ is only a function of the inner image gradient vector. The second transformation P₂ is a function of both the inner and outer image gradient vectors. For the hypothesis H₁ (target), the difference of the two functions is small. For the hypothesis H₀ (clutter), the difference of the two functions is large. Results of testing the proposed target detection algorithm on two large FLIR image databases are presented.

The spectral angle mapper (SAM) has been widely used in multispectral and hyperspectral image analysis to measure spectral similarity between substance signatures for material identification. It has been shown that the SAM is essentially the Euclidean distance when the spectral angle is small. Most recently, a stochastic measure, called the spectral information divergence (SID), has been suggested to model the spectrum of a hyperspectral image pixel as a probability distribution, so that spectral variations among spectral bands can be captured more effectively in a stochastic manner. This paper develops a new hyperspectral spectral discrimination measure, which combines the SID and the SAM into a mixed measure. More specifically, let r and r′ denote two hyperspectral image pixel vectors with their corresponding spectra specified by s and s′. Then SAM(s,s′) measures the spectral angle between s and s′. Similarly, SID(s,s′) measures the information divergence between the probability distributions generated by s and s. The proposed new measure, referred to as the SID-SAM mixed measure, can be implemented in two versions, given by SID(s,s′)×tan(SAM(s,s′)) and SID(s,s′)×sin(SAM(s,s′)), where tan and sin are the usual trigonometric functions. The spectral discriminability of such a mixed measure is greatly enhanced by multiplying the spectral abilities of the two measures. In order to demonstrate its utility, a comparative study is conducted among the SID-SAM mixed measure, the SID, and the SAM. Our experimental results have shown that the discriminatory ability of the (SID,SAM) mixed measure can be a significant improvement over the SID and SAM.

A novel technique is proposed for data fusion of earth remote sensing. The method is developed for land cover classification based on fusion of remote sensing images of the same scene collected from multiple sources. It presents a framework for fusion of multisource remote sensing images, which consists of two algorithms, referred to as the greedy modular eigenspace (GME) and the feature scale uniformity transformation (FSUT). The GME method is designed to extract features by a simple and efficient GME feature module, while the FSUT is performed to fuse most correlated features from different data sources. Finally, an optimal positive Boolean function based multiclass classifier is further developed for classification. It utilizes the positive and negative sample learning ability of the minimum classification error criteria to improve classification accuracy. The performance of the proposed method is evaluated by fusing MODIS/ASTER airborne simulator (MASTER) images and the airborne synthetic aperture radar (SAR) images for land cover classification during the PacRim II campaign. Experimental results demonstrate that the proposed fusion approach is an effective method for land cover classification in earth remote sensing, and improves the precision of image classification significantly compared to conventional single source classification.

A new synthetic hit-miss transform (SHMT) algorithm is proposed for distortion-invariant recognition of various objects in noisy and cluttered input images. The proposed SHMT algorithm uses synthetic structuring elements (SEs), which are synthesized based on a synthetic discriminant function (SDF) filter algorithm. The synthetic hit SE is composed of the linear combination of the reference hit SEs, and the synthetic miss SE is composed of the linear combination of the reference miss SEs. Based on various simulations, it is shown that the proposed algorithm can be applied to an HMT correlator to improve its ability to detect various objects with distortions in noisy and cluttered scenes.

Face recognition with both pose and illumination variations is considered. In addition, we consider the ability of the classifier to reject nonmember or imposter face inputs; most prior work has not addressed this. A new SVRDM (support-vector representation and discrimination machine) classifier is proposed, and initial face recognition-rejection results are presented.

A new method for fingerprint identification is proposed, using an ANFIS-based matching algorithm, which is suitable for large-scale identification systems. In this algorithm, the Gabor transform is used to extract features of fingerprints, while ANFIS is trained to identify the resultant Gabor features. Experimental verifications show that this proposed matching algorithm has high accuracy.

Biometric authentication can provide an added level of security and/or ease of convenience in access control applications. Fingerprints are a popular choice among the biometric features and have been successfully used in criminal identification. In access control applications, we are interested in obtaining digital live-scan fingerprints from sensors, rather than the inked fingerprints usually used in criminal identification. In this paper, we evaluate the performance of composite correlation filters in fingerprint verification for access control applications. The NIST Special Database 24, obtained from an optical fingerprint sensor, is used to evaluate the performance of fingerprint verification in the presence of distortion.

There are an increasing number of high-occupancy-vehicle lanes throughout the world. In these lanes each vehicle is required to have two or more occupants. These laws are currently manually enforced. This paper presents a new system for automating the enforcement process, using color image segmentation techniques to find the windshield area, and face detection to count the number of occupants within a vehicle.

We propose a correlation-based digital watermarking technique for robust image pattern authentication. We hide a phase-based signature of the image back into its Fourier magnitude spectrum in the embedding stage. The detector computes the Fourier transform of the watermarked image and extracts the embedded signature. Authentication performance is measured by a correlation test of the extracted signature and the signature computed from the watermarked image. The quality of the watermarked image is obtained from the peak signal-to-noise ratio metric. We also furnish simulation results to show the robustness of our approach to typical image processing as found in JPEG compression

We propose a hybrid filter, which we call the hybrid optical neural network (HONN) filter. This filter combines the optical implementation and shift invariance of correlator-type filters with the nonlinear superposition capabilities of artificial neural network methods. The filter demonstrates good performance in maintaining high-quality correlation responses and resistance to clutter to nontraining in-class images at orientations intermediate to the training set poses. We present the design and implementation of the HONN filter architecture and assess its object recognition performance in clutter.

Depth detection and image recovery of a target from its complex hologram have recently been proposed in optical remote sensing applications. The proposed technique involves synthesizing the so-called real-only spectrum hologram from the complex hologram. The resulting spectrum hologram contains the real-only information of the target's spectrum and carries the target's depth information. By digital processing of the spectrum hologram, we can extract the target's depth. Once the depth is obtained, we can recover the image of the target from its complex hologram. We report an experimental verification of the technique by optical acquisition of the complex hologram of the remote target and subsequently extract the target's depth. The image of the target at its depth location is subsequently recovered digitally. The novelty of this remote sensing technique is that by performing a single 2-D scan, we can acquire simultaneously the target's depth information and its image for possible subsequent optical correlation applications.

For practical pattern recognition and tracking systems, it is often useful to have a high-speed random access memory (RAM) that complements a holographic correlator. Recently, we have demonstrated a super-parallel holographic correlator, which uniquely identifies N images from a database using only O(N^½) number of detector elements. We show how this correlator architecture, operated in reverse, may be used to realize a super-parallel holographic random access memory. We present preliminary results, establishing the feasibility of the superparallel holographic random access memory, and show that essentially the same set of hardware can be operated either as the super-parallel holographic optical correlator or as a super-parallel holographic random access memory, with minor reorientation of some of the elements in real time. This hybrid device thus eliminates the need for a separate random access memory for a holographic correlator-based target recognition and tracking system.

We propose and demonstrate a photorefractive real-time holographic deconvolution technique for adaptive one-way image transmission through aberrating media. In contrast with preceding methods, which have typically required various coding of the exact phase or two-way image transmission for correcting phase distortion, our technique relies on one-way image transmission using exact phase information. Our technique can simultaneously correct both amplitude and phase distortions. We demonstrate our results through both experiment and computer simulation for different aberrators.

This paper describes a system for tracking objects in a video stream obtained from a moving airborne platform. The system is applied to annotate video objects automatically. The object to annotate is indicated by a mouse click. The proposed tracking algorithm is based on the general framework of spatiotemporal segmentation followed by temporal tracking. However, it differs from existing techniques in the following features. The same algorithm is used in tracking both moving and stationary objects by making the stationary objects "move." It is general enough to handle objects of various types and sizes, including point objects. The system has a fast implementation because all image operations are applied to small image regions only. We demonstrate the effectiveness of the proposed algorithm on a few real video sequences.

Ultrafast temporal pattern generation and recognition with femtosecond laser technology is presented, analyzed, and experimentally implemented. Ultrafast temporal pattern generation and recognition are realized by taking advantage of two well-known techniques: the space-time conversion technique and the ultrafast pulse measurement technique. Here the temporal pattern for the designed multiple pulses, optimized with a preassumed Gaussian spectral distribution of an ultrashort pulse, is described. With the simulation of a Gaussian spectral distribution, we realize that the uniformity of the generated multiple ultrafast temporal pulses is relevant to the repeated number of modulation periods in the mask in the spectral plane. Moreover, the change of Gaussian spectral phases with the wavelengths in the modulated phase plate is considered. Experiments of ultrafast temporal pattern recognition by the frequency-resolved optical gating (FROG) characterization technique are also given.

A two-stage algorithm is proposed for locating smooth and detailed disparity vector fields in a stereo image pair. The algorithm consists of hierarchical disparity estimation using a region-dividing technique and edge-preserving regularization. The hierarchical region-dividing disparity estimation increases the efficiency and reliability of the estimation process. At the second stage, the vector fields are regularized with an energy model that produces smooth fields while preserving discontinuities resulting from object boundaries. The minimization problem is addressed by solving a corresponding partial differential equation using a finite-difference method. Experiments show that the proposed algorithm provides accurate and spatially correlated disparity vector fields in various types of stereo images, even in the case of images with large displacements.

Texture is one of many important features to capture in image characteristics. A recent texture unit-based texture spectrum approach, referred to as texture unit coding (TUC) developed by Wang and He has shown promise in texture classification. We present a new texture feature extraction coding, called gradient texture unit coding (GTUC) that is based on Wang and He's texture unit to capture gradient changes in a texture unit. Since the GTUC also generates an 8-D ternary texture feature vector in the same way that the TUC does, a GTUC-generated feature vector can be further represented by a number in the same range generated by the TUC. As a result, the GTUC-generated numbers also form a texture spectrum similar to that formed by the TUC-generated numbers. By normalizing a texture spectrum as a probability distribution, this work further develops an information divergence (ID)-based discrimination criterion to measure the discrepancy between two texture spectra, a concept yet to be explored in texture analysis. To compare the GTUC to the TUC in texture classification, several criteria used in hyperspectral image analysis are also introduced for performance analysis.

The Fourier transform method is applied to the input complex function of a nonlinear system characterized by a transfer characteristic, where the real and imaginary parts of the input function undergo either distinct or the same nonlinear transformations. The effects of the complex transfer characteristic at the output of the system are investigated. As an example, an analytical expression describing the complex output of a symmetrical odd-power-law complex nonlinear system is provided.

Aliasing is unavoidable in real-time computer image generation due to the sampling processes occurring within the graphics hardware. In particular, aliasing produces scintillation effects and significant radiometric inaccuracy when targets are rendered at long range. This problem was alleviated some years ago by the development of a zoom anti-aliasing (ZAA) technique within infrared missile seeker hardware-in-the-loop simulations. An alternative ZAA technique based on extensive use of available graphics hardware functions is described here and compared to the original technique.

A newly developed fiber Bragg grating (FBG) fabrication method is demonstrated. By combining a stable continuous-wave UV laser, a high precision electro-optical UV modulator, a compact and stable holographic setup, and a subnanometer precision air-bearing translation stage, the setup allows both apodization and phase of the FBGs to be continuously changed at each grating line. During the writing process, a fiber is moved constantly by the high-precision air-bearing stage, while the UV laser beam is switched on and off by a customized pulse train with nanosecond timing precision and forms a stable interference pattern on the fiber. The error of this system mainly results from the control signal's timing jitter, which is normally lower than 0.1%. Theoretical analysis and experimental results for gratings with various structures are demonstrated and discussed. They show good agreement, demonstrating that the fabrication method is capable of fabricating gratings with arbitrary apodization and phase design.

We describe the fabrication and characterization of ridge structure on Z-cut LiNbO₃ substrates. Titanium strip patterns are formed using conventional photolithography and are used as a mask. The unmasked portions are then proton exchanged. As proton-exchanged LiNbO₃ have a much higher etch rate than bulk LiNbO₃, these portions are etched away easily by a suitable etchant, leaving a ridge structure along the masked portion. Different characterizations have been done both before and after annealing the proton exchanged areas. These ridge structures can be used for making broadband modulators.

Most telecom components use lenses to collimate the beam emerging from a single-mode fiber and again, after some manipulation of the beam, to focus the beam back into an output fiber. In response to the drive for smaller components, collimator lenses have shrunk and a number of new microcollimator types been introduced. Several of the newer collimator types involve lenses that are fused to the fiber. Optical software was used to model six different types of microcollimators divided into three different size categories. Because not all of the six types were applicable in all three of the size categories, a total of eleven different collimator systems were modeled. A collimator system consists of an input collimator and an identical output collimator. Each collimator includes a single-mode fiber and a collimating lens. In order to achieve equivalence, all of the collimators in the same size category produce an intermediate waist with the same mode field diameter and distance from the lens surface to the intermediate waist. An algorithm is presented to determine the necessary physical parameters (lengths and radii) for a collimator to produce the appropriate beam criteria for a given size category. Once the physical parameters for the collimator systems were determined, all of the collimator systems were examined for insertion loss in their ideal configuration as well as in the presence of fabrication and alignment errors. All of the insertion loss analyses were performed using coherent field propagation starting from the fundamental fiber mode of the input fiber, through the collimating lenses, and finishing with the calculation of the overlap integral of the field inside the output fiber with the fundamental fiber mode. No thin lens or paraxial propagation assumptions are made...The results indicate that the collimator types involving lenses that are fused to the fiber are less sensitive to longitudinal fiber and collimator shifts than the components with the fiber and lens not in direct contact.

All-optical switches are used for many different applications. Broadcast-enhanced switches allow connection of a single input channel to numerous output channels, thus increasing further the possible application span. A compact multichannel all-optical broadcast-enhanced switch is described. This switch setup allows electronic alignment thus requires only simple construction procedures. A special importance of the proposed switch is for multistage interconnection networks where the alignment plays a major role. The switch structure is described and experimentally demonstrated.