In classical microscopy and inverse scattering, images are made from far field measurements. It will be shown that the standard Abbe or Rayleigh resolution limits arise as a result of assumptions which do not hold in the near field. Existing technologies such as TIRM and PSTM that exploit these properties of the near field produce two dimensional images which exhibit structure on a subwavelength scale. These images have proven difficult to interpret when the sample presents variations in structure and optical properties simultaneously in three dimensions. We present the analytic singular value decomposition of the linearized scattering operator. This representation leads to a robust inversion algorithm for the inverse scattering problem in the near zone. Numerical implementation of the algorithm is demonstrated using simulated noisy data to produce tomographic images of a three dimensional scatterer.

A novel design of a simple reflection confocal microscope with a multimode graded-index fiber output as an effective way for improving the dynamic range of the resolving power and achieving a submicron spatial resolution is presented. The basic idea for obtaining high spatial resolution involves the following advanced scheme properties. (1) It is an apertureless confocal arrangement in which the input laser emission with its maximum power is launched directly to the test object. (2) It uses a high-numerical-aperture focusing objective (NA>0.8) that provides high depth and spatial discriminations and thus, high axial and lateral resolutions are obtained. (3) It uses a graded-index multimode fiber (50-μm-core diameter) for signal detection rather than a conventional pinhole or single-mode fiber. The graded-index fiber ensures efficient launching and maintenance of high laser powers, and high sensitivity of signal detection. Moreover, according to the equivalent-Gaussian-mode model, the intensity mode distribution into the graded-index fiber can be presented as an equivalent Gaussian mode distribution. As a result, a significant reduction of the mode noise of the intensity distribution and of signal fluctuations is obtained. Thus, a regime of high-output power is achieved that provides high resolving power and a submicron axial resolution of 0.4 μm.

Polarization control and analysis continue to be a useful tools for microscopy. We have previously addressed the use of spatially inhomogeneous polarization, specifically azimuthal polarization, in scanning optical microscopy. In this paper we describe new techniques for generating both azimuthally and radially polarized cylindrical-vector beams, investigate high numerical aperture focusing, compare the theoretical and experimental point spread functions for small particle imaging, and explore the use of these beams with microscopy.

he quadrature tomographic microscope combines conventional microscopy with interferometric quadrature detection techniques to yield a simultaneous measurement of both the amplitude and phase of the electric field. This is of tremendous utility in imaging nearly-transparent object's since changes in phase can be directly attributed to changes in the objects index of refraction while the small amplitude changes provide additional structural information. In this paper we will discuss the progress we have made towards reconstructing three dimensional images from the measured two-dimensional complex electric field. By applying an artificial curvature to the measured field and back-propagating the Gaussian beams, we are able to digitally manipulate the imaging kernel and image different focal planes.

We present a programmable array microscope (PAM) which uses a pair of ferroelectric liquid crystal spatial light modulators (FLC SLMs). The system is similar to a confocal imaging system called a direct-view microscope (DVM), in which scanned aperture arrays are used to obtain real-time confocal images. The PAM, unlike the DVM, allows arrays to be scanned electronically rather than mechanically. In our system, one SLM is placed in the source plane of a conventional microscope system; the other is placed in the detector plane. Confocal aperture arrays are displayed and scanned synchronously on the SLMs and confocal imaging results. The resolution is improved when compared to a similar previously presented system which employed a single SLM. We present axial resolution measurements for a variety of array dimensions and investigate the use of aperture correlation techniques to improve the light throughput in the devices.

This paper discusses some aspects of standardization in digital microscopy. Included are image acquisition , use of image and graphics formats and data protocols in multidimensions, particular image analysis procedures and use of industry standards for software development

The functional significance of dendritic spines and their plasticity to a wide spectrum of developmental and pathological conditions has led to extensive studies based on spine morphology. The advances in image acquisition techniques and the associated generation of large 3D data sets of optical micrographs have not been accompanied by comparable advances in data analysis techniques. We present an automated 3D spine detection and quantification procedure suitable for images obtained by laser scanning microscopy. The image is first processed by deconvolution and the dendritic phase consisting of the neuronal cytoplasm is extracted by segmentation. Spines are detected as geometrical protrusions relative to the dendritic backbone. As very thin necks may not be imaged, some spine `heads' may be detached from the dendrite and are detected as detached components. These detected heads are merged with spine `bases' where appropriate. Morphological characterizations on spine length, volume, density and shape classifications are obtained. For time-lapse data, images are registered and individual spines are traced through the image sequence. Successful comparison results on spine lengths and densities with manual analysis are obtained. This method is highly automatic and allows detailed and objective quantification of the structure and dynamics of dendritic spines, which can be important predictors for the function of neural networks.

The three-dimensional imaging properties of a light microscope are traditionally described through an intensity point spread function (PSF) or its Fourier transform, the optical transfer function (OTF). However, the imaging properties can be more compactly described by a generalized two-dimensional pupil function. Use of the pupil function allows easy modification of an observed PSF to introduce known aberrations, a much more difficult task when using a PSF or OTF. Unfortunately, it is not straightforward to determine the complex-valued pupil function from the measured intensity PSF, because of the lack of phase information. This is the problem of phase retrieval. Several phase retrieval algorithms have been developed for two-dimensional imaging in astronomy. We have modified one such algorithm to be appropriate for the high-aperture, non-paraxial case of high resolution light microscopy. PSFs reconstructed from phase-retrieved pupil functions, modified by calculated aberrations closely reproduce the features of measured, aberrated PSFs.

With the advent of laser scanning confocal and multiphoton microscopy, 3D life tissue characterization has been rendered possible. This involves the restoration of thick section images (in the depth range of 100 microns) of biological samples. In contrast to thin samples new effects become important when imaging thick samples: Because of changes of the refractive index across the specimen or the embedding medium, strong (>1 micron) focal shifts and spherical aberration occur and scattering effects get more prominent. For tissue mapping it is essential to correct for such aberrations and distortion effects. In this paper, we propose a calibration framework, which allows us to determine thick section focal shifts and spherical aberrations in a one-step-procedure. Gradients in the focal shift induce a scaling in the Z-direction of the observed sample. A second effect arises with depth-dependent spherical aberration. We model our microscope as a linear-shift-NON-invariant system (LSNI) where multiple depth classes are assigned distinct point spread functions (PSFs). We measure these two effects in a 3D sample of randomly distributed fluorescent focal check beads. The beads are embedded in gelatine, a medium with high resemblance to real biological tissue. The PSF is approximated by a mathematical parametric model. While estimating the parameters of the PSF with object-constrained deconvolution, we track depth dependent changes of the observed bead diameter. This allows us to determine the gradient of the focal shift across a thick section. By numerical integration along the optical axis we obtain the focal shift values as a function of sample depth. In the end, these values will be applied as a correction to compensate depth distortions in the tissue images.

We used a new type of polarized light microscope (Pol-Scope) equipped with a precision universal compensator to record images of small, uniaxial birefringent crystals. Pol-Scope measurements of small calcite crystals represent complete in-focus and out-of-focus retardance maps of the crystals. We relate the 3-D retardance maps to the crystal morphologies and the directions of their optic axes in three dimensional space. In addition to single crystal measurements, we also recorded 3-D retardance maps of small clusters of crystals to study the superposition of overlapping image features. We demonstrate that the retardance of overlapping image features in cluster images can be theoretically predicted by combining separately recorded single crystal images. Our experimental and theoretical analysis indicates that the superposition of retardance features is equivalent to adding incoherently the respective intensity values that were recorded with the Pol-Scope using different compensator settings. This analysis is part of a larger effort towards establishing 3-D image restoration procedures for polarized light imaging.

Focused Ion Beam (FIB) is a novel technique that allows easy target cell selection, fast operation, high resolution, 3D imaging and sample manipulation during imaging. The FIB technique of microscopy and nanomachining, that is widely spread in semiconductor technology, is reshuffled to open new horizons in the field oflife sciences at cellular and subcellular level. FIB is a source ofions which can be precisely oriented and focused on the sample, supported also by an electron source. The resolution can be as low as I 5nm. Two different operation modes are available: "cutting" and "etching" operations of very high precision and at the same time the resulting secondary ions and electrons provide follow up sample imaging. FIB tomography ability has been tested to provide information on cell characterisation, cell division time sequence, view of inner structures and investigation of membrane structural properties. We compare the performances and the advantages of different high-resolution microscopy techniques: Soft X-ray Contact Microscopy (SXCM), Focused Ion Beam (FIB) and Transmission Electron Microscopy (TEM). These have been used to image Saccharoniyces cerevisiae yeast cells, a rather well known sample of great biological interest. TEM is an established technique used as a benchmark for comparison.

We present a study of the polarization effects resulting from the reflection and transmission of a narrow beam of light through biological tissues. This is done numerically with a Monte Carlo method based on a transport equation which takes into account polarization of light. We will show both time-independent and time-dependent computations, and we will discuss how polarization can be used in order to obtain better images. We consider biological tissues that can be modeled as continuous media varying randomly in space, containing inhomogeneities with no sharp boundaries.

A compact excitation-emission spectrofluorimeter system used to acquire tissue fluorescence and reflectance with fiber optic light delivery and collection is presented. Ten dye cells are pumped by a XeC1 excimer laser, creating laser excitation from 308 nm to 505nm, which is used to excite tissue fluorescence. A xenon-flash lamp is employed to obtain the reflectance spectrum of the corresponding tissue spot. Eleven fluorescence spectra and one white light diffuse reflectance spectrum are collected in vivo in 0.2 seconds and can be potentially analyzed in real time to obtain diagnostic information. The instrument can be used for early cancer detection. The measured fluorescence and reflectance spectra are used to model the intrinsic fluorescence of the tissue, which gives important insight into its biochemical composition (tryptophan, collagen, elastin, NADH, etc.) and metabolic state. Diffuse reflectance provides information about tissue morphology, in terms of the scatterhig and absorption coefficient. By subtracting the diffusely reflected component from the measured reflectance, light scattering spectroscopic (LSS) information due to single backscattering from epithelial cell nuclei can be obtained. LSS provides information about the size distribution of cell nuclei. The parameters give diagnostic information, which can be used to guide the physician in real time while taking biopsies of invisible tissue abnormalities.

The objective of this work is to develop endoscopic subsurface optical imaging technology that will be able to image different tissue components located underneath the surface of the tissue at an imaging depth of up to 1 centimeter. This effort is based on the utilization of existing technology and components developed for medical endoscopes with the incorporation of the appropriate modifications to implement the spectral and polarization difference imaging technique. This subsurface imaging technique employs polarization and spectral light discrimination in combination with image processing to remove a large portion of the image information from the outer layers of the tissue which leads to enhancement of the contrast and image quality of subsurface tissue structures.

Tissue-like phantom experiments were performed in order to investigate the possibility of using the polarized light scattering spectroscopy for skin cancer diagnosis. Using a two-layer model, particle size distribution (PSD) was analyzed for a thin layer of latex spheres on top of 1% Intralipid. The results show that PSD of the latex spheres in the top layer can be extracted with and without a priori information about the shape of the particle size distribution.

We have developed a noninvasive confocal fluorescence spectrometer to measure in vivo fluorescence spectra in mice. Our motivation is the study of the healing process of wounded cartilage in the murine model. The spectrometer measures the expression of green fluorescent protein (GFP) which is linked to the promoter of type II collagen as a marker for the expression of the chondrocyte phenotype. The confocal system uses an argon ion laser (488 nm) to excite fluorescence in a confocal volume about 200 um below the skin surface in the cartilage of the xyphoid process. In vitro studies showed how the refractive mismatch at the surface boundary and the light scattering of the tissue affect the depth and size of the confocal volume. A thin fluorescent layer was detected through a 130 um tissue phantom (murine skin) by moving the confocal volume axially through the phantom. The presence of skin increased the axial full width at half maximum of the confocal response from 7 um to 72 um. In conclusion, we achieve a low-spatial-resolution confocal spectrometer, which yields a fluorescence spectrum that maximizes the GFP fluorescence of cartilage and minimizes the skin autofluorescence.

This paper investigates the behavior of a Faraday-active turbid media under ultrafast optical excitation. As the degree of polarization of early arriving photons is mostly preserved during the first 100 ps after the arrival of the ballistic component, it is addressed the possibility of using the magneto-optical rotation of the light polarization as a tool for tissue characterization. The technique may be useful to determining photon scattering statistics in turbid biological media. The analysis is performed on the basis that photons trapped in multiple scattering events leave the medium with larger induced rotation angles. It was found that the experimental observation of picosecond magneto-optical phenomena in scattering media of biological origin is submitted to constraints in the applied magnetic field intensity and in the statistical parameters of photon scattering in the turbid media.

Native fluorescence spectra of tissues of both normal and cervical carcinoma were measured at 320 nm excitation. The spectra were normalized with respect to maximum peak intensity and average fluorescence spectra were computed for normal and cancerous cervical tissues. The average fluorescence spectrum of normal cervix is found to be significantly different from that of cancer. Both normal and cancerous cervical tissues show a distinct primary emission peak around 380 nm which may be attributed to the structural proteins, collagen and elastin. In addition, the average spectrum of normal cervix shows a less intense peak around 440 nm, which is red shifted to 460nm and more intense in the case of cancerous cervical tissues, when compared to normal. A ratio parameter I380/I460 is introduced and a critical value of 1.9 for this ratio parameter is found to classify cervical cancerous tissues from normal with a sensitivity and specificity of 100% and 85.7% respectively. Also the slope value analysis between 420 to 440nm & 440 to 460nm of the normal and cancerous samples gives 100% accuracy in the classification of normal from the cancerous samples. Further, discriminant analysis together with Principal component analysis (PCA) was performed to determine the statistical significance and to confirm the above two analysis. The potentiality of the ratio parameter and slope values analysis is improved by the support of multivariate statistical analysis.