Traditional approaches in confocal microscopy have focussed on techniques that use elastically, or Rayleigh, scattered photons to generate volumetric intensity or phase images of an object. Common to these imaging modes is an inability to discriminate between optically similar but chemically distinct materials. We report in this paper on a new class of confocal microscope which uses inelastically, or Raman, scattered light to generate volumetric chemical images of a material. We designed and built a prototype instrument, called a confocal scanning laser FT-Raman microscope, which combines a confocal scanning laser microscope with a FT Raman spectrometer. The high depth and lateral spatial resolution of the confocal optics design defines a volume element from which the Raman scattered light is collected and then analyzed by the spectrometer for its spectral content. The sample is scanned through the microscope focal volume and a 3D chemical image is generated based on the content of the Raman spectrum measured at each scan position. The results to be presented include instrument characterization measurements and examples of volumetric chemical imaging.

We will present the design of a greatly simplified acousto-optic (A/O) scanning system which allows a change of wavelength in less than a second (and in principal between 2 TV lines). A/O deflectors now available with a 9.3 mm circular entrance pupil (rather than the 2 mm X 12 mm pupils previously used) eliminate the need for costly anamorphic beam shaping optics. The resulting simplified optic system can be straightforwardly corrected for different excitation wavelengths.

We demonstrate how the new Intensity-modulated Multiple-beam Scanning technique can be used in improve simultaneous recording of multiple fluorophores in confocal scanning laser microscopes.

A mirror deflection device for a CSLM has been developed. It performs repetitive scanning according to a preset waveform which can be chosen arbitrarily. It can also be used to perform stationary positioning at arbitrarily chosen points. A digital memory, comprising dual banks, is used to allow switching from one actuating waveform to another. The movement of the mirror is recorded very accurately. A burst of sequential pulse from a diode laser is deflected by the mirror and recorded by means of a linear diode array. The target pattern is analyzed digitally. The objective is to implement a control strategy whereby a new actuating waveform can be derived in order to correct any deviation between the desired waveform and the recorded one. Some results obtained with the device are reported. Foreseen applications encompass spectral analysis of selected regions and kinetic studies where a trade-off between speed and number of image points is necessary.

The size and weight of conventional optical microscopes often makes them inconvenient for use on the human body or for in-situ examination during materials processing. We describe a new fiber-optic scanning confocal optical microscope which could have a total outside diameter as small as 1 mm, and should lend itself to applications in endoscopy or to optical in vivo histology. The first experimental device utilizes a single-mode optical fiber for illumination and detection. The scanning element is a mechanically resonant fused silica cantilever 1.5 cm long and 0.8 mm across, with a micromachined two-phase zone plate objective mounted at one end. The cantilever is electrostatically scanned near resonance in two dimensions, generating a Lissajous pattern which is scan converted to conventional video for real time display or digitization. The objective lens has N.A. equals 0.25 at (lambda) equals 0.6328 micrometers , with a measured spot size of 1.8 micrometers FWHM.

We have designed and constructed an experimental confocal specimen-scanning microscope which has the capability of producing high resolution 3D images in a variety of optical modes, many of which are not currently available on commercial confocal microscopes. The transmission Nomarski differential interference contrast mode is particularly interesting because it can be utilized to image small changes in refractive index within complex biological specimens which are transparent in standard brightfield. The three-color reflection configuration can produce a color 3D image, which means that stained or pigmented objects will be similar in appearance to images obtained from conventional white light microscopes which makes them more recognizable.

The aim of this paper is to explore the implications of a multiplicative model of optical imaging.

Resonant two photon excitation of UV absorbing dyes with red light is a new concept in 3D microscopy. This technique provides functionality which is virtually equivalent to a confocal microscope. Two photon excitation allows excitation of UV dyes without special UV corrected optics. Resonant two photon excitation in microscopy, as known until now, utilizes expensive, high powered pulsed femto- or picosecond lasers. In this paper a cost effective alternative to resonant two photon is presented.

We demonstrate the possibility to study, simultaneously, the variations over the image area of the lifetimes of two different fluorophores, in a confocal laser microscope. In this way information conveyed by the lifetimes can be extracted, in particular complementary information from two different fluorophores. The fluorophores are excited by laser light of two different wavelengths, which are modulated at different frequencies using electro-optical modulators. By frequency-selective detection, using two lock-in amplifiers, it is possible to efficiently separate signals that emanate from each of the fluorophores. Further, by using 2- phase lock-in amplifiers the phase of each of the separated signals is determined. Since the phase shift of the emitted light relative to the exciting light depends on the lifetime of the fluorophore, the technique allows mapping of the lifetime. Two images are obtained, one for each fluorophore. The method can be extended to supply information concerning the excitation cross sections at the two laser wavelengths used.

A system for 3D lifetime measurements of fluorophores has been developed and applied to multiple stained specimens. The system is built around a picosecond detector system for time correlated single photon counting and a confocal scanning laser microscope. It is capable of measuring lifetimes in the interval 1 - 20 ns, with a resolution of approximately 10 ps. By studying the lifetime of the fluorescence in a specimen it is possible to identify different fluorophores even if their emission spectra overlap. In addition, information regarding the chemical environment may be obtained due to effects from the environment on the fluorescence lifetime of the fluorophore.

Real-time confocal imaging is subject to a number of constraints connected with the emission capabilities of (especially) fluorescent specimen and the particular confocal imaging technique employed. We will see that there are from the confocal image collection techniques no basic impediments towards real-time 3D imaging. The limitations sooner lie in the specimen fluorescence emission capabilities--both with respect to emission rate and total emission and of course biological tolerance limits to the exciting radiation. The various factors are examined and it is found that parallel confocal illumination and detection approaches especially via techniques of direct field (also known as direct view) confocal imaging. These combined with detection on CCD detectors offer probably the optimal and most convenient way to realize real-time imaging.

The confocal imaging of thin film structures is investigated. A theoretical treatment of imaging of stratified media with continuously varying refractive index is presented, and the inverse problem of reconstructing the refractive index profile from a confocal image discussed.

This paper presents a technique for producing accurate surface profile measurements using optical differentiation and confocal microscopy. An experimental system based on this technique has been constructed for the purpose of surface profiling. Experimental results are presented and compared to measurements from other systems. Theoretical transfer functions are also presented to characterize system performance.

Imaging properties of an optical system may be modified by the introduction of spatial filters at the entrance and exit pupils. A classic example of this is the annular pupil, which is known to improve lateral resolution at the expense of depth discrimination. In confocal fluorescence microscopy, previous studies have shown that a configuration with an annular pupil in the excitation path and ca circular pupil in the emission path offer little improvement. We show here that a more favorable situation results if the annular excitation pupil is design with a partially transmissive central obstruction. Our simulations show that the radii and leakage of the annulus may be adjusted to improve the point spread function's lateral resolution by 13 percent and with not degradation of the Z-axis full width at half-maximum. A slightly different pupil is shown to improve the lateral resolution by 10 percent and the Z-axis response by 5 percent.

Scanning optical microscopy computer assisted has the advantages that a wide range of imaging modes is possible. We have constructed a new computer assisted laser scanning which offers the possibility to obtain 2D and 3D images. The paper presents the principal hardware and software characteristics of the scanner.

This work aims to study in vivo vesicle formation, transport and fusion during the digestive process of the ciliated protozoan Paramecium primaurelia.

Four-dimensional (4D) microscopy, the acquisition of time-resolved 3D images of living biological samples, provides not only structural data but also contains information on the mechanics and forces inside a living cell. In order to extract, quantify, and visualize such information for chromosomes, we are exploring simple computational methods that will allow quantitative estimation of chromosome motion within 4D images of living nuclei. A correspondence-based approach is required due to limited temporal resolution. The chromosomes we are tracking are relatively textureless, but additional structural constraints can be imposed to discriminate against false matches.

Hyaline cartilage is composed of chondrocytes that reside in lacunae surrounded by extracellular matrix molecules. Microscopic and histochemical features of cartilage have been studied with many techniques. Many of these techniques can be time consuming and may alter natural cartilage characteristics. In addition, the orientation and order of sectioned tissue must be maintained to create 3D reconstructions. We show that confocal laser scanning microscopy may replace traditional methods for studying cartilage.

Real-time confocal scanning microscopy has many desirable features. Living specimens which are moving rapidly can only be imaged with a real-time confocal microscope. A new real- time, scanning slit, confocal microscope designed for the clinical in vivo examination of the human cornea and anterior segment is described. The key feature of the in vivo confocal microscope is the use of two conjugate adjustable slits, and a two-sided oscillating mirror which is used both for scanning and for descanning.

Even in confocal scanning, longitudinal resolution is poorer than lateral resolution. It is therefore of interest to go `beyond confocal' and achieve still better optical sectioning by image restoration methods. In our previous work we applied two methods to simulated 3D microscope images: the constrained Jansson-van Cittert (JVC) method, which is a deterministic regularized image restoration algorithm, and the expectation-maximization (EM) algorithm, which is a method to obtain the maximum likelihood solution of the restoration problem under Poisson image statistics. In this paper we apply both the JVC algorithm and the EM algorithm to real image data obtained from our laser scanning confocal microscope. Slices of the original and restored images agree with our earlier numerical simulations. Specifically: (a) optical sectioning is improved by both algorithms; (b) the JVC restoration is noisier than the image restored with the EM algorithm, showing the advantage of the ML approach under low light conditions; (c) noise in the EM restoration shows that regularization is still needed.

Confocal theta microscopy is a novel method by which the axial resolution in confocal fluorescence microscopy can be substantially improved. The basic idea is to observe the sample with two or more objective lenses and to detect the emission light at an angle theta to the illumination axis. The observation volume is considerably decreased when the detection axis is rotated by 90 degree(s) relative to the illumination axis. This leads to an almost spherical observation volume, which is three times smaller than in a comparable confocal fluorescence microscope. Confocal theta microscopy can be combined with 4Pi-confocal microscopy and is the first viable method proposed for fully exploiting the resolution increase achievable with 4Pi-techniques. The axial side lobes of the 4Pi-point spread function are suppressed, and the observation volume is reduced by a factor of 4. The resolution properties of confocal theta fluorescence microscopies are investigated for single- and two-photon absorption. Evaluations of confocal theta point spread functions are presented and the resolution improvement achieved by theta observation is discussed.

This paper discusses how digital image quality criteria help to optimize image quality, in particular for applications in laser scanning microscopy. Image quality considerations offer a uniform description of the available transfer characteristics, which are summed up and weighted properly to finally represent the system by a single number. In the spatial domain we can measure sharpness and contrast of the (digital) volumes by analyzing intensities and their local dependencies in a statistical fashion. This includes sum modulus difference, gray level variance, and lateral inhibition. Based on information theory, the criterion volume fidelity takes into account the knowledge of the spatial structure of a test object and compares the intensities with those present in the final digital image. Applications presented here include measurement of image quality improvement when going from non-confocal to confocal imaging, testing of new confocal system designs and the evaluation of digital post-processing methods. Limitations in the presence of noise are discussed.

Nicer-Slicer-Dicer (NSD) is a new, interactive tool for volume visualization. In addition to single scalar volumes, NSD can visualize multivariate and time series data. NSD achieves high rendering rates by compositing slice planes back-to-front while blending pixels through the use of alpha buffering hardware. This technique introduces some small artifacts, but it works remarkably well with large data sets where voxels represent small areas in image space. The plane compositing technique also allows one to generate an isosurface and volume rendering simultaneously by interleaving planes and isosurface polygons.

The topics addressed in this paper include (1) computer reconstruction of the capillary tree of the human renal glomerulus from serial sections, (2) computer visualizations for the purpose of analyzing the resulting 3D structures, and (3) computer quantification of these structures using 3D morphology. The purpose of the research reported is to develop methodology that will permit better understanding of the physical structure of the human renal glomerulus and, in a more general manner, the structure of 3D biological networks embedded in volumetric data.

We describe a new interaction technique based on head-coupling designed to free a user's hands for other tasks such as 3D tracing. We used head motion to manipulate the 3D view of confocal microscope data of neurons. The simplest interaction mode shows the projected view from the user's eye point. Other modes implement more sophisticated motions such as ratcheting. Tracing is done by marking a point of interest in two different views controlled by the head. Under suitable restrictions this locates a point in 3D corresponding to a feature of interest in the data set. By repeatedly marking points in this way the user traces a 3D path through the data. Because of pointing errors and rendering artifacts, the path traced by a user needs to be refined by consulting the original data set. Using an rough data model, along with several successive points in a feature, the program can automatically supply the next point thereby implementing a form of semi-automatic tracing. An advantage of our method is that we have access to the entire 3D volume data, so feature geometry can be computed with reference to the complete 3D data and not just to a 2D projected view.

The purpose of the present article is to describe the reconstruction of 3D histological structures and of their spatial configurations by using a set of images obtained from histological serial sections acquired with a conventional microscope. A sequence of image processing algorithms is proposed and utilized for the reconstruction of a global slice from partial views, for the extraction of the contours of objects of interest, for a geometric registration of image planes, and for the final visualization of the spatial configurations of such objects.

This paper presents a new, non-destructive technique for high resolution 3D measurements of microstructures based on digital photogrammetry.

Axial resolution in fluorescence microscopy can be improved significantly by using standing wave illumination to selectively excite planes within the depth of field of the microscope. When the specimen is thinner than 0.18 micrometers , an estimate of its 3D structure may be determined from three images within the same focal plane without re-focusing. Thicker objects require a combination of multi-focal-plane data and/or a priori knowledge.

Our initial system for 3D reconstruction of neural tissue from transmission electron microscope (TEM) images has been improved and expanded in functionality and scope. An automated acquisition system captures images of tissue and controls the movement of a TEM. These images comprise a dataset of roughly one gigabyte. Using these data, software running on a Connection Machine automatically reassembles individual images into a single image of each section. An automated contour extraction and object classification algorithm is used and the objects to be reconstructed are selected by the user. Registration is completely automated, but the result is user verifiable and modifiable. The registration parameters are then used to realign both the contour and raw image data. The contour data are smoothed to average out noise, a surface grid is generated, and the resulting reconstruction is visualized. The image data can also be volume visualized. The result is a completely digital, easy-to-use, quantifiable, and generalizable system for 3D reconstruction from transmission electron microscope serial sections.