Biomedical subjects

T Pun

Publications and source records attributed to T Pun.

11 recordsLinked to original sources

PubMed1992

Comparison of photodensitometric with high-resolution digital analysis of bone density from serial dental radiographs.

Photodensitometry is known to provide high spatial resolution and continuous measurement of optical density for the analysis of dental radiographs, whereas digitization allows powerful image manipulations but, when using conventional video cameras, gives less spatial resolution and fewer grey levels. The aim of this study was therefore to develop a technique of high-resolution digital analysis for the measurement of bone density following the same principles as those of photodensitometry and based upon the use of a CCD Scanner Camera which provides up to 4096 grey levels and a spatial resolution of 4096 x 4096 pixels. Twenty-four zones were analysed with both techniques in five serial dental radiographs taken before and after periodontal therapy in eight patients. Statistical comparison of the results obtained by digital analysis and photodensitometry shows that the two techniques have the same accuracy.

Absorptiometry, Photon↗

PubMed1991

LaboImage: a workstation environment for research in image processing and analysis.

Numerous images are produced daily in biomedical research. In order to extract relevant and useful results, various processing and analysis steps are mandatory. The present paper describes a new, powerful and user-friendly image analysis system: LaboImage. In addition to standard image processing modules. LaboImage also contains various specialized tools. These multiple processing modules and tools are first introduced. A one-dimensional gel analysis method is then described. The new concept of 'normalized virtual one-dimensional gel' is introduced, making comparisons between gels particularly easy. This normalized gel is obtained by compensating for the bending of the lanes automatically; no information loss is incurred in the process. Finally, the model of interaction in a multi-window environment is discussed. LaboImage is designed to run in two ways: interactively, using menus and panels; and in batch mode by means of user-defined macros. Examples are given to illustrate the potentialities of the software.

Algorithms↗

PubMed1990

From biopsy to automatic diagnosis.

High resolution two-dimensional gel electrophoresis is a very powerful biochemical tool for analysis of complex protein mixtures. In well defined situations, protein maps, obtained from tissue biopsies or biological fluids by this technique, can be automatically analyzed by computer. Some polypeptide patterns are the fingerprints of diseases. Applying clustering algorithm and learning techniques, the prototype expert system MELANIE recognized patterns and associated the correct diagnosis to the specific pattern.

Diagnosis, Computer-Assisted↗

PubMed1988

"High-resolution" mini-two-dimensional gel electrophoresis automatically run and stained in less than 6 h with small, ready-to-use slab gels.

Although two-dimensional (2-D) gel electrophoresis is one of the most powerful techniques for analyzing protein mixtures, its application in routine clinical laboratories is currently limited, because it is time-consuming, complex, and relatively expensive. Here we describe a method for automatically running and staining "high-resolution" mini 2-D electrophoresis gels in less than 6 h, by using "ready-to-use" slab gels and a PhastSystem electrophoresis apparatus. We present 2-D gel electrophoretograms of 25 nL of plasma, as well as their automatic computer analysis. For comparison, a conventional 2-D gel electrophoresis profile of 200 nL of a plasma sample is shown. The technique is easy to perform, highly sensitive, rapid, and potentially useful in semi-routine clinical chemistry laboratories.

Autoanalysis↗

PubMed1988

Computerized classification of two-dimensional gel electrophoretograms by correspondence analysis and ascendant hierarchical clustering.

A powerful data processing methodology for analysis and classification of two-dimensional gels is introduced. The approach is based on correspondence analysis (CA) and ascendant hierarchical classification (AHC), and significantly differs from the more classical principal-component decomposition. Starting with a series of gels, each having a large number of spots, CA allows their representation in a factorial space of reduced dimension; classification into meaningful groups is then performed using AHC. Simultaneous representation of both spots and gels in the same space can be done. This precisely indicates the key spots pertinent for the classification, and therefore the characteristic proteins representative of a particular class of gels (i.e. of a particular disease or biological status). In addition, knowledge of these characteristic spots greatly simplifies the screening of future gels. After a brief overview of the Mélanie system for analyzing 2D gels, the theory of correspondence analysis and ascendant hierarchical classification is summarized. Equations are given that are easily ammenable to computation. How classification of two-dimensional gel electrophoretograms is accomplished is then detailed. Experimental results support the power of this approach.

Cluster Analysis↗

PubMed1987

Optimal background estimation in EELS.

In quantitative electron energy loss spectrometry, it is desirable to estimate the background law below core edge energy in a way that provides the maximum signal-to-noise ratio. Assuming an inverse power background model and independently Poisson distributed measurements, it is shown how to achieve this goal by using a maximum likelihood (ML) estimation technique which provides unbiased and minimum mean square error estimates of all parameters of interest. An efficient and computationally stable implementation of this procedure is proposed. Standard logarithmic least squares estimations are then compared with the ML approach and the gain in performance due to optimal processing is quantified.

Electrons↗

PubMed1986

Hexavalent capsomers of herpes simplex virus type 2: symmetry, shape, dimensions, and oligomeric status.

The structures of the hexavalent capsomers of herpes simplex virus type 2 were analyzed by negative staining electron microscopy of capsomer patches derived from partially disrupted nucleocapsids. Optimally computer-averaged images were formed for each of the three classes of capsomer distinguished by their respective positions on the surface of the icosahedral capsid with a triangulation number of 16; in projection, each capsomer exhibited unequivocal sixfold symmetry. According to correspondence analysis of our set of capsomer images, no significant structural differences were detected among the three classes of capsomers, as visualized under these conditions. Taking into account information from images of freeze-dried, platinum-shadowed nucleocapsid fragments, it was established that each hexavalent capsomer is a hexamer of the 155-kilodalton major capsid protein. The capsomer has the form of a sixfold hollow cone approximately 12 nm in diameter and approximately 15 nm in depth, whose axial channel tapers in width from the outside towards the inner capsid surface.

Capsid↗

PubMed1985

Weighted least squares estimation of background in EELS imaging.

In quantitative Electron Energy Loss Spectrometry, a weighted least squares estimation should theoretically be used to estimate the background law below core edge energy, since the variances of the data vary. However, it is found that proper weighting makes the above edge signal-to-noise ratio decrease rather than increase. This result is discussed, and the influence of the bias introduced by the logarithmic transformation of the data is quantified.

Spectrum Analysis↗

PubMedDate not supplied

Automatic learning strategies and their application to electrophoresis analysis.

Automatic learning plays an important role in image analysis and pattern recognition. A taxonomy of automatic learning strategies is presented; this categorization is based on the amount of inferences the learning element must perform to bridge the gap between environmental and system knowledge representation level. Four main categories are identified and described: rote learning, learning by deduction, learning by induction, and learning by analogy. An application of learning by induction to medical image analysis is then exposed. It consists in the classification of two-dimensional gel electrophoretograms into meaningful distinct classes, as well in their conceptual description.

Electrophoresis, Gel, Two-Dimensional↗

PubMedDate not supplied

An expert system for guiding image segmentation.

Application of image segmentation to biomedical research is now customary. Due to the existence of a rich heuristic knowledge, many users have no deep experience in this field. It is therefore necessary to integrate knowledge-based techniques with image segmentation operators. The purpose of our expert system is to guide users in image segmentation. Its main functions are: suggest a reasonable overall scheme of processing and recommend appropriate operators and algorithms at each stage. The characteristics of this expert system are presented: (a) interaction with users: through "conversation," the expert system acquires the informations about a given problem; (b) use of belief values which indicate users' descriptions about the image characteristics. By associating image features with belief values, the system gets the informations about the image appearance and makes inference more effectively; (c) local backtracking strategy, which allows the expert system to repeatedly search for a better solution until a satisfactory result is obtained; (d) integrating with an image analysis package, users can directly execute the operations recommended by the expert system. A practical application of the system is then shown in details. Finally, our opinions in designing such a system are discussed.

Algorithms↗

PubMedDate not supplied

Colour displays and look-up tables: real time modification of digital images.

Image processing in biomedical research has become customary, along with use of colour displays to run image processing packages. The performance of softwares is highly dependent on the device they run on: architecture of colour display, depth of frame buffer, existence of look-up table, etc. Knowledge of such basic features is therefore becoming very important, especially because results can differ from device to device. This introductory paper discusses hardware features and software applications. A general architecture of colour displays is exposed, comparing the features of the most commonly used devices. Basic organisation of memory, electron gun and screen are analysed for each type of display, concluding with a more detailed study of raster scan devices. Frame buffer and look-up table organisation are then analysed in relation with overhead expenses such as time and memory. Relation between image data and displayed images is discussed. By means of examples, the manipulation of colour tables is examined in detail, showing how to improve display of images without altering image data. Finally, the basic operations performed by the look-up table editor developed at University of Geneva are presented.

Color↗