An analysis is presented of the image formation in widefield fluorescence microscopy using standard light. The region of support of the resulting optical transfer function is discussed.

Although scattering for spheres with plane wave illumination was solved precisely by Mie in 1909, often it is of interest to image spheres with non-planar illumination. An extension of Mie theory which incorporates non-planar illumination requires knowledge of the coefficients for a spherical harmonic expansion of the incident wavefront about the center of the sphere. These coefficients have been determined for a few special cases, such as Gaussian beams, which have a relatively simple model. Using a vectorized Huygen's principle, a general vector wavefront can be represented as a superposition of dipole sources. We have computed the spherical wave function expansion coefficients of an arbitrarily placed dipole and hence the scattering from a sphere illuminated by a general wavefront can be computed. As a special case, Mie's solution of plane wave scattering was recovered. POtential applications include scattering with partially coherent illumination. Experimental results from the scattering from polystyrene spheres using Koehler illumination show agreement with numerical tests.

A multi-anode photomultiplier tube (PMT) attached to a light microscope was used to measure the low-light-level signals from fluorescent dyes trapped in the extracellular space between living, cultured kidney epithelial cells. The detection assembly utilized photon counting of all 64 channels to record the fluorescence produced after the photoactivation of caged fluorophores of different molecular weight and charge. Photoactivation was accomplished by a brief pulse of light at 355 nm forma frequency-tripled Nd:YAG laser. Fluorescence of the uncaged fluorophores was excited by the output of an argon laser equipped with an acousto-optical tunable filter for control of wavelength and power. Diffusion coefficients for the fluorescent indicators in the extracellular spaces were calculated from the spatial and temporal decay of the fluorescence after uncaging of the dye in a small region micrometers -diameter microscope field. The system magnification was adjusted so that each 2.5-mm-square PMT channel corresponded to a 12.5-micrometers square region of the microscope field. The spatial decay of the fluorescence was obtained by sampling multiple adjacent PMT channels, while the temporal decay was determined from the PMT channel encompassing the uncaging site.

In this paper we summarize the imaging possibilities and performance offered by the Intensity-modulated Multiple- wavelength Scanning (IMS) technique. The IMS technique is based on intensity-modulated laser illumination of the specimen in combination with lock-in detection of the fluorescent light. We have used this technique with multiple illumination wavelengths, spectral separation of the fluorescent light, and multiple detectors. It is then possible to substantially extend the imaging possibilities of confocal fluorescence microscopy. Among the new possibilities are: 1) improved spectral separation when recording the light intensities of multiple fluorophores, 2) separation of intensity signals from multiple fluorophores based on lifetime differences, 3) simultaneous lifetime recording of multiple fluorophores. We have investigated how different system parameters influence the channel separation and signal-to-noise ratio (SNR). In order to maximize the SNR, factors such as fluorophore lifetimes and modulation frequency of the laser light are important. In our investigations we have obtained good agreement between theoretical and experimental results. In summary, our results show that when using the IMS technique to scan multiple-fluorophore specimens, very high channel separation can be obtained for both lifetime and intensity images. For lifetime images, however, the SNR is worse by a factor of approximately four compared with intensity images.

Blind deconvolution microscopy, the simultaneous estimation of the specimen function and the point spread function (PSF) of the microscope is an under-determined problem with non- unique solutions. The non-uniqueness is commonly avoided by enforcing constraints on both the specimen function and the PSF, such as non-negativity and band limitation. These constraints are sometimes enforced in ad hoc ways. In addition, many of the existing methods for blind deconvolution estimate the PSF pixel by pixel thus greatly increasing the number of parameters to estimate the slowing the convergence of the algorithm. We derived a maximum- likelihood-based method for blind deconvolution in which we assume that the PSF follows a mathematical expression that depends on a small number of parameters. The algorithm then estimates the unknown parameters together with the specimen function. The mathematical model ensures that all the constraints of the PSF are satisfied and the maximum likelihood approach ensures that the specimen is non- negative. This parametric blind deconvolution method successfully removes out-of-focus blur but its degree of success depends on the features of the specimen. Specimen features that fall in mostly the null space of the PSF are more difficult to recover and make PSF estimation more difficult.

Imaging thick specimens in 3D transmission confocal modes presents two key problems. The first problem is variable aberrations introduced by changes in refractive index. The second problem is revealed when visualizing acquired data, where thick 3D datasets are difficult to interpret. In this paper we present our emerging solutions to these problems. Aberrations can be classified as simple tip-tilt deflection of the beam, or more complicated higher order aberrations. We discuss our results which demonstrate successful on-the- fly detection and correction for tip-tilt. For detecting higher order aberrations, we have chosen to investigate the wavefront curvature sensing technique. The second problem of rendering thick 3D datasets can be solved by extracting features of interest from the background. Simple intensity thresholding is not sufficient for complex biological specimens. And image processing in only 2D neglects any 3D structure. Use of Kohonen's self-organizing map neural network in 3D results in clear segmentation of features for sample chromosome specimens.

Two different phase-estimation methods that have been developed for the computation of the optical path-length (OPL) distribution of a specimen from DIC image for the determination of the OPL distribution. The second phase- estimation method is based on the conjugate-gradient optimization method and estimates the OPL distribution using rotational-diversity DIC images; i.e. multiple DIC images obtained by rotating the specimen. For this study, 24 different DIC images of a single bovine spermatozoa head acquired by rotating the cell by approximately 15 degrees between images. The images were registered and aligned using fiducial marks, and then processed with both methods. Results obtained with the filtering method were found to be dependent on the orientation of the cell with respect to the shear direction. Comparison of the integrated optical path length (IOPL) computed with the filtering method and the rotational-diversity method using two, four and eight DIC images at different rotation angels showed that the IOPL estimated with the rotational-diversity method is less dependent on the rotation angel, even when only two images separated by 90-degree cell rotation are used for the phase estimation. Our results show that the use of rotational- diversity images in the determination of the OPL distribution is very beneficial because it overcomes the directional dependence of DIC imaging.

In this work we discuss the role of polarization in confocal microscopy. The effect of the optical system on light polarization is analyzed in the absence of a specimen. We also present numerical results on confocal microscopes imaging small dielectric scatterers and suggest an optimum pinhole size for crossed linear polars. Finally, imaging of large spherical scatterers and contrast optimization of the optical system is discussed.

A deconvolution algorithm for use with scanning laser ophthalmoscope (SLO) data is being developed. The SLO is fundamentally a confocal microscope in which the objective lens is the human ocular lens. 3D data is collected by raster scanning to form images at different depths in retinal and choroidal layers. In this way, 3D anatomy may be imaged and stored as a series of optical sections.Given the poor optical quality of the human lens and random eye motion during data acquisition, any deconvolution method applied to SLO data must be able to account for distortions present in the observed data. The algorithm presented compensates for image warping and frame-to-frame displacement due to random eye motion, smearing along the optic axis, sensor saturation, and other problems. A preprocessing step is first used to compensate for frame-to-frame image displacement. The image warping, caused by random eye motion during raster scanning, is corrected. Finally, a maximum likelihood based blind deconvolution algorithm is used to correct severe blurring along the optic axis. The blind deconvolution algorithm contains an iterative search for subpixel displacements remaining after image warping and frame-to-frame displacements are corrected. This iterative search is formulated to ensure that the likelihood functional is non-decreasing.

In previous studies, we examined 3D images of a latex bead containing a surface layer of fluorescence whose thickness was determined by physical sectioning. Confocal images of the bead from six different microscopes all exhibited a significantly thicker fluorescence shell than actually present. In contrast, deconvolved wide-field images of the bead produced an image with the correct shell thickness. We have now repeated some of these studies using a new latex bed containing a much thinner layer of surface fluorescence. In contrast to earlier studies, confocal image of this bead appear to show an accurate thickness for the fluorescence shell. These particular confocal microscopes also yielded essentially aberration free images, which was not the case for some of the earlier microscopes tested. In these recent studies however, deconvolution of the bead appeared less robust than earlier. Of three independent wide-field images of the bead, only one yielded a substantially artifact-free image upon deconvolution.

3D detailed visualization of neurons represents a source of valuable information when trying to understand integrative phenomena at cellular and circuital levels. We describe here a low cost 3D reconstruction system developed by an interdisciplinary team of engineers and neurobiologists. The system allows 3D reconstruction of ultra thin sections observed through a transmission electron microscope. We have developed and tested a completely automatic registration method which combines local histogram equalization with correlation in multi-resolution. This method gives good experimental results in real images. It is not time consuming and it does not need dedicated hardware. The different parts of the system are briefly described: image acquisition, non-uniform illumination compensation, manual segmentation, automatic registration, 3D visualization and graphic user interface. Preliminary experimental results are also included here.

We introduce an optical means of reconstructing the topology of rough objects under the microscope or at a relatively high magnification. First, 2D images were obtained under the conventional optical microscope or under the macroscopic imaging system, focused on differing height of the sample. The depth information was extracted by sensing the focus of each pixel of the conventional image slice. Each sample height associated with the best 'focus figure' for each pixel was marked and were later utilized to generate the 3D coordinate map.

A cooled or video-rate CCD camera is often used to collect optical slices from many modalities of microscopy including widefield-fluorescence (WF) and transmitted-light brightfield (TBL). Raw optical slices collected from both of these types of cameras are contaminated by imperfect performance of the CCD camera. Optical slice image data must be corrected for the bias level and nonuniformity in the photometric response of CCD elements. Fluctuation of the exposure time in a cooled CCD camera needs to be calibrated as well. Bad pixels due to the severely low sensitivity of some of the CCD elements often occur and otherwise impair the conventional schemes of correction and calibration. An adaptive median filter has been introduced by us previously to treat these bad pixels. In this study a Wiener-type regularization scheme is proposed as robust, fast, and practical alternative treatment of these bad pixels. One advantage of using this scheme is that doing so provides a nicely modularized design of the computer algorithm. As such, the correction scheme is software readily integrated as a software component in a 3D microscopy system and may operate independently without affecting the design of other software components in the system.

The aim of this work is to show that properly trained ANNs can be used as an image restoration tool to correct the effect of defocusing on 2D optical microscopy images. The proposed method can be applied to correct the results of inaccurate range focusing algorithms on fully automated imaging and analysis systems and to compensate the effect of the limited of the depth of focusing on high numerical aperture system. One type of ANN was used: feedforward multilayer perceptron, with supervised back propagation training. Its performance has been tested with both synthetic images and real images from latex microspheres. The network was trained with sets of pairs of images: each pair consisted of a defocused image and its corresponding in-focus version. Different levels of defocusing were used. The criteria used to select the algorithm parameters to tune the networks and to train them will be presented. The results of the experiments performed to test their ability to 'learn' to correct the defocusing and to generalize the results will also be shown. The result show that, when trained with images with some levels of defocusing, the network was able to learn and accurately correct these defocusing levels, but it can also generalize the results and correct other levels of defocusing.

This paper describes a method for the improvement of biological images acquired using a transmission electronic microscope. Several techniques are presented that deal with noise reduction, artifact removal and non-uniform illumination correction. Experimental results are shown.

The automatic reconstruction of 3D structures from stacks of 2D images is an important problem in medical image analysis. In neuroscience, in particular, the availability of explicit 3D models of the dendritic tree structure of a neuron can be a valuable tool for understanding the complexity of neuronal function and neuronal morphology In this work, the dendritic tree structure is imaged at different levels through the tissue using a laser scanning confocal microscope. The aim of the software is to deliver an explicit representation of the tree as a generalized cylinder model. Automatic techniques often fail to produce a continuous 3D model and manual or semi-automatic techniques can be particularly labor intensive.In this paper we describe algorithms which lead to a continuous 3D cylinder model representation of the dendritic tree with a minimum of user involvement. The initial stage of the approach involves the identification of voxels with a high probability of being on the center-lines of the dendritic structure. These points are linked to form a skeleton using cost minimization techniques. In the final stage the full 3D structure is extracted around the center- lines and represented by generalized cylinders. This paper provides details of the algorithms in each of the stages together with some of the results of using the method on both synthetic and real data sets.

Tomographic measurements of the 3D refractive index spatial distribution within optically transparent phase samples with computerized interferometric microscopes are proposed. Phase shifting interferometric microtomography applications for the 3D image reconstruction of the blood cells are represented. The immersion 100x, N.A. equals 1.25 objective was used to increase the spatial resolution and observation angle range to 90 degree. ART, combined ART and iterative Gerchberg-Papoulis 3D algorithm were used for the tomogram reconstruction. To determine the accuracy and spatial resolution of the blood cells image reconstruction by means of the interferometric microtomographic method the numerical simulations were implemented.

Confocal microscopy is qualified to perform volume scans of nerve cells with dendrites and spines. Length and diameter of the dendrite branches and the spines should be determined to analyze the influence of learning processes. A prerequisite for that is the recognition of the dendritic structure with branching off and spines. Because the microscope operates at the resolution limit the images are blurry, noisy and only poorly sampled. In contrast to other methods which are based on binary images and thinning algorithms, our method tracks and dendritic tree faster and in the gray-level domain using simple geometric models.An explicit segmentation is unnecessary and knowledge about shape and structure of the dendrite is included as a priori information. For large trees, first a low resolution scan is captured to crete a rough model. The algorithm allows to refine this model using higher resolution scans for interesting regions along the dendrite. The large unimportant areas between the dendrite branches are not scanned at high resolution to save time and disc space. In a second step, the parameters of the model are adapted to the microscope image by minim zing the deviation of the microscope image from the mode image convolved by the microscope point spread function. Features like number, diameter, length and position of the dendrite branches and spines can be easily calculated from the model. An interactive user intervention is possible at the model domain.

Image acquisition at high magnification is inevitably correlated with a limited view over the entire tissue section. To overcome this limitation we designed software for multiple image-stack acquisition (3D-MISA) in confocal laser scanning microscopy (CLSM). The system consists of a 4 channel Leica CLSM equipped with a high resolution z- scanning stage mounted on a xy-monitorized stage. The 3D- MISA software is implemented into the microscope scanning software and uses the microscope settings for the movements of the xy-stage. It allows storage and recall of 70 xyz- positions and the automatic 3D-scanning of image arrays between selected xyz-coordinates. The number of images within one array is limited only by the amount of disk space or memory available. Although for most applications the accuracy of the xy-scanning stage is sufficient for a precise alignment of tiled views, the software provides the possibility of an adjustable overlap between two image stacks by shifting the moving steps of the xy-scanning stage. After scanning a tiled image gallery of the extended focus-images of each channel will be displayed on a graphic monitor. In addition, a tiled image gallery of individual focal planes can be created. In summary, the 3D-MISA allows 3D-image acquisition of coherent regions in combination with high resolution of single images.

In our lab we have developed an atomic force microscope (AFM) for biological and technical applications with improved optical method for measuring of cantilever with an additional lens placed between cantilever and four-segment photodiode. The lens forms an image of the cantilever in the photodiode plane. Small sizes of the cantilever image improve the resolution of AFM and reduce the requirements to the optical scheme of the microscope. The lens nonlinear transmission of the cantilever deflection is compensated by means of computer program. With the AFM were imaged the DNA of plague microbes and phages of plague and V. Cholera.

A computational model of the image formation process has been developed for the Nomarski differential interference contrast (DIC) microscope. The DIC microscope images variations of the phase of the light wave transmitted through the specimen. In the study of biological phenomena, the DIC microscope is used to visualize live cells which are highly transparent in the visible spectra but distort the phase of the impinging light wave. Within the microscope, a birefringent prism splits the transmitted light wave into two laterally sheared wavefronts. An interference pattern is imaged when the wavefronts recombine. The computational model we developed uses polarization ray-tracing techniques. Rays propagating through different microscope components and the specimen are traced. A specimen is represented by a 3D grid of voxels, each containing a complex refractive index. At the image plane, a coherent superposition of the diffracted field due to each ray contributes to the image intensity. Partial coherence at the image plane is also simulated by wavefronts with different propagation directions. By computing the image intensity at different positions along the axial direction, we can obtain optically sectioned images. In order to evaluate our model, we compared simulated images to the images taken under a real DIC microscope. We constructed test specimens of known shape and properties, using polystyrene beads in optical cement and an etched glass wafer. As the next step, we plan to use this computational model for the reverse problem, i.e. to reconstruct the 3D refractive index distribution of an imaged specimen.

An IR spectroscopic imaging technique has been developed which combines a step-scan Fourier transform (FT) Michelson interferometer with focal plane array (FPA) image detection. The first reports utilized an indium antimonide FPA in the spectral region 3950-1975 cm^-1. Mercury-cadmium- telluride (MCT) focal plane arrays are now commercially available, giving access to the spectral range 3950-800 cm^-1, thus greatly broadening the application of the technique for chemical analysis in the mid IR region. This paper describes some of the instrumental considerations as well as applications in the use of an MCT array detector, for novel mid-IR spectroscopic imaging applications.

The overwhelming size of a hyperspectral image creates serious problem for users to understand and utilize the data. A novel unsupervised neural network (UNN) model is presented. The UNN is designed to analyze the spectral contents of the multi-spectral images. The UNN automatically grows its layers and neurons by scanning the training images and by learning spectral features. The learning strategy is optimized to ensure fast convergence. At the end of learning, the UNN provides a table of the spectral content of the images. The image contents are categorized based on spectral similarities. The resulting spectral classes are then mapped onto the image, thus the UNN effectively compresses a hyperspectral image cube into a single image. The UNN is also able to automatically recognize objects by their spectral features.the ability of the UNN in identifying subtle spectral differences is shown.

The goal of imaging spectroscopy is to obtain independent spectra from individual objects in a field-of-view. In the case of biological materials, such as histopathology samples, it has been well established that spectral characteristic can be indicative of specific diseases including cancer. Diagnosis can be enhanced by the use of probes and stains to indicate the presence of individual genome or other biologically active cell components or substances. To assess a specimen through a microscope is directly analogous to serving the Earth from space to assess natural features. This paper describes a simple and inexpensive imaging spectrometer, with an origin in remote sensing, that demonstrates that it is possible to rapidly identify evidence of disease in histopathology samples using spatially resolved spectral data. The PARISS imaging spectrometer enables a researcher to acquire multi-spectral images that yield functional maps, showing what and where biological molecules are located within a structure. It is the powerful combination of imaging and spectroscopy that provides the tools not readily available to the Life Sciences. The PARISS system incorporates a powerful hybrid neural network analysis to break the data logjam that is often associated with the acquisition and processing of multiple spectra.

A novel optical approach to predicting chemical and physical properties based on principal component analysis (PCA) is proposed and evaluated using a data set from earlier work. In our approach, a regression vector produced by PCA is designed into the structure of a set of paired optical filters. Light passing through the paired filters produces an analog detector signal directly proportional to the chemical/physical property for which the regression vector was designed. This simple optical computational method for predictive spectroscopy is evaluated in several ways, using the example data for numeric simulation. First, we evaluate the sensitivity of the method to various types of spectroscopy errors commonly encountered, and find the method to have the same susceptibilities toward error as standard methods. Second, we use propagation of errors to determine the effects of detector noise on the predictive power of the method, finding the optical computation approach to have a large multiplex advantage over conventional methods. Third, we use two different design approaches to the construction of the paired filter set for the example measurement to evaluate manufacturability, finding that adequate methods exist to design appropriate optical devices. Fourth, we numerically simulate the predictive errors introduced by design errors in the paired filters, finding that predictive errors are not increased over conventional methods. Fifth, we consider how the performance of the method is affected by light intensities that are not linearly related to chemical composition, and find that the method is only marginally affected. In summary, we conclude that many types of predictive measurements based upon use of regression vectors and linear mathematics can be performed more rapidly, more effectively, and at considerably lower cost by the proposed optical computation method than by traditional dispersive or interferometric instrumentation. Although our simulations have used Raman experimental data, the method is equally applicable to NIR, UV-Vis, IR, fluorescence and other spectroscopies.

Different kinds of cell surface receptor clusters have been discovered recently using fluorescence resonance energy transfer (FRET) measurements. This method is capable for identifying molecular interactions, however the exact distances remain obscure, because the classical Foerster efficiency-distance relationship is valid only in the case of one donor one acceptor systems. This condition can not be fulfilled when cell surface molecules are labeled with monoclonal antibodies carrying different number of fluorescent donor and acceptor molecules. Our aim was to carry out FRET measurements on such cell surface receptors, where the distances are constant, and the only changing parameter is the donor-acceptor ratio of the used labels. For our experiments we used JY B lymphoblastoid cells, and we labeled the MHC class I heavy chain with KE-2 or W6/32 monoclonal antibodies and the length chain with L-368 monoclonal antibody tagged with different numbers of donor or acceptor molecules. The FRET efficiencies were measured either in a microscope using the photobleaching method or in a fluorescence activated cell sorter. We changed the donor acceptor ratio in a wide range in order to make a suitable calibration curve for other FRET experiments. The obtained calibration curve gives us the possibility to relate FRET efficiencies to real distances of cell surface receptors. Another source of deviation in the FRET efficiencies arise from the selected method. There was a marked difference between the FRET efficiencies measured by flow cytometry and with the photobleaching method even on same cells and between same epitopes.

Recent improvements in filters, multi-element detectors and instrument design have transformed Raman spectroscopy from a difficult to use specialist technique into a widely used multi-dimensional spectroscopic method. Raman spectroscopy is non destructive and offers a spatial resolution of one micro or better. A Raman spectrum gives specific information regarding the chemical bonding of molecules and can therefore be used to identify different molecules in a system. Through the use of xyz mapping techniques, specific types of material can be imaged in living cells, drug formulations and polymer mixtures to give but a few examples. Raman technologies allow areas as large as 500 microns to be imaged directly using filters tuned specifically to look for a particular chemical species. The Raman technique uses visible or close to visible light which is ideal for coupling into optical fibers. It is therefore very easy to build ruggedized spectrometers using fiber optic probes for remote sensing in extremely difficult and/or hazardous environments; for example process monitoring and recently endoscopic diagnostic work in living subjects. This paper will describe the methodology used in direct Raman imaging, Raman mapping experiments and remote sensing with reference to specific examples of biological, pharmaceutical, mineral and crystal studies.

Laser induced fluorescence is widely being developed as a new method for early detection of pre-cancerous tissue. Typically an optical fiber probe directly contacting tissue is used for detection and localization of abnormal tissue. This method of tissue interrogation has certain clinical and diagnostic restrictions due to physical contact with the tissue. This method of tissue interrogation has certain clinical and diagnostic restrictions due to physical contact with the tissue. The objective of our study was to investigate the method of remote sensing in colposcopy. The device was designed using an optical multichannel analyzer with UV excitation. The excitation light is delivered through the fiber bundle to the optical head that is attached to the colposcope. This arrangement allows one to conveniently collect fluorescence spectra from the cervix during regular colposcopy examination. To understand the practical restrictions of the remote sensing the influence of various factors are studied. A few modifications of the device were tested. Phantoms were created and used to test the device. Data form more than 200 patients shows substantial improvement in sensitivity and specificity compared with the doctor's impressions. The result of the experimental measurements of tissue-like phantoms and in- vivo measurements are provided. The modification increased the average signal/noise ration more than 4 times.

Multi-spectral images have been the basic source of information in various fields of science and technology for acquiring the multi-spectral images. After introducing a basic and simple technique of parallel Fourier spectrometry, we present the principle and experimental results of parallel techniques of ultra-fast spectral imaging. The spatially resolved interferometric data are simultaneously detected by the parallel optical systems realized by an array of small lenses. Introduced are two equivalent optical configurations realized by using a conventional Michelson interferometer and a compact interferometer incorporating wedged liquid-crystal cells. These spectral imagers have been applied to fast spectral-imaging experiments: spectral imaging of a rapidly rotating object and investigation of laser ablation of liquid. Also addressed is the coherence- based spectro-tomographic system that aims to detect the 3D spatial distribution and also the spectral distribution of scattered light in turbid media. We adopted a parallel layout of confocal optical system to exclude scattered light form the interference signals.

A novel optical approach to single-shot chemical imaging with high spectroscopic resolution is described using a prototype dimension-reduction fiber-optic array. Images are focused onto a 30 X 20 array of hexagonally packed 250- micrometers o.d. f/2 optical fibers which are drawn into a 600 X 1 distal array with specific ordering. The 600 X 1 side of the array is imaged with an f/2 spectrograph equipped with a holographic grating and a CCD images and deconvolute them into wavelength-specific reconstructed images or position-specific spectra which span a 190-nm wavelength space. 'White-light' zero-order images and first- order spectroscopic images of laser plumes have been reconstructed to illustrate proof-of-principle.

Surface vegetation is an important link in the coupling between the atmosphere and the biosphere. Monitoring the condition of vegetation cover on the Earth surface is essential for detecting the changes in climate. Advanced Very-High Resolution Radiometer 10-day composite data in 1 X 1 degree resolution from NASA/GSFC and a global vegetation ground truth in the same resolution from the University of Maryland's Geography Department are used in this study. A fully connected multilayer neural network is used for supervised classification. The normalized difference vegetation index, which is also called the greenness index, is used along with the surface reflectance and brightness temperature as the input features. Trainings and classifications are performed for two spatial modes and three multitemporal modes.

Optical imaging of tissue with low-coherence interferometry offers the potential of non-invasive imaging and the possibility of spectral analysis of tissue characteristics. We suggest an interferometric technique that allows us to determine the optical properties of reflective boundaries and backscattering sites in a scattering medium. The sample is placed in ore arm of a Michelson interferometer. The interferogram fringes are formed only when the difference between the optical paths in two arms is less than the coherence length. The positions of peaks in the interferogram provide the information on locations of reflective boundaries and backscattering sites. Simultaneously, the spectral properties about the scattering sites are retrieved by filtering the interferogram with various band-pass filters. We propose a low-coherence confocal interferometer in combination with a microlens array for parallel scanning and processing. The system has the capability of rejecting scattered light from out-of- focus planes.As a low-coherence light source, we use the white light continuum, which is generated by focusing amplified femtosecond pulses into a water containing cell. The white light continuum has the broad spectrum and high brightness.

Hyperspectral Raman microscopic imaging of carbonated hydroxyapatite (HAP) is used to follow the chemistry of bone growth and regrowth. Deep red excitation is employed to minimize protein fluorescence interference. A passive line generator based on Powell lens optics and a motorized translation stage provide the imaging capabilities. Raman image contrast is generated from several lines of the HAP Raman spectrum, primarily the PO₄^-3. Factor analysis is used to minimize the integration time needed for acceptable contrast and to explore the chemical species within the bone. Bone age is visualized as variations in image intensity. High definition, high resolution images of newly formed bone and mature bone are compared qualitatively. The technique is currently under evaluation for study of experimental therapies for fracture repair.

A common path interferometric element introduced in the optical path of an imaging device is a well documented method to perform multidimensional spectroscopy. Recent design modifications however have provided significant improvements including enhanced spectral resolution and optical throughput, reduced acquisition time, as well as reduced instrument weight and volume. The new design will be reviewed in addition to its impact on three applications: spectral karyotyping, spectral imaging of the human ocular fundus and remote sensing of water reservoirs.

We describe a method of obtaining optically scanned fluorescence images in a widefield conventional microscope by interfering two beams on an object so as to illuminate it with a single spatial frequency fringe pattern. Images taken at three spatial positions of the fringe pattern are processed in real time to produce optically sectioned images which are substantially similar to those obtained with confocal microscopes.

The relationship between chemical composition and microstructure is becoming increasingly more important in industry as a way of identifying key performance related properties of fully formulated products. Raman micro- spectroscopic imaging has become the preferred method of obtaining this type of information due to the wealth of detail inherent in Raman spectra pertaining to both microstructure and chemical composition. Concurrently, rapid chemometric data elucidation methodology is required to be able to simultaneously extract pertinent chemical and concentration information in a self-modeling manner from these typically huge and complex data sets. Examples of Raman micro-spectroscopic imaging will be presented as a demonstration of how the principles of Raman imaging can be applied to complex, multi-component, multi-phase systems of inherently low contrast. The use of two-way multivariate curve resolution (MCR) methodology is described as a means of rapidly processing the extremely large, three-way Raman chemical images, dramatically simplifying data analysis. The results from MCR analysis provide the number of chemical species present in the sample, the spectrum of each species for identification, and the concentration image for each species. The additional benefit of image noise reduction is also described for the MCR approach. A comparison of results is given describing the two-way PFA methods currently in use and how they compare with MCR methodology.