Patrizia Ciarlini

Some Basic Questions in the Treatment of Intercomparison Data

An analysis of the interlaboratory comparison process highlights the need to pose some basic questions that should be answered before the exercises take place. They often remain hidden in the protocol while proper statistical tools should be adopted to process the data provided by the N participants in different cases. Intercomparisons are here subdivided into different classes according to the probability model type that represents the stochastic variability of the involved data. This paper also introduces the use of a mixture probability model to characterize the temperature data supplied in temperature intercomparison as a single population. In this approach, the reference value of the intercomparison becomes the natural definition of the expected value of this mixture probability function.

Application of special reduction procedures to metrological data

To improve the reduction of metrological data, that are typically grouped in series and cannot be considered as replicated data, a modelling procedure has been obtained by adding to the model representing the physical behaviour, common to all data, a specific term for each series. Such a procedure combines both the advantages of preserving the individuality of each series and of improving the variance estimate which arises from fitting the overall data. A non-parametric bootstrap method for the error analysis has been developed, which does not imply the assumption of the Normal distribution in the least squares estimation. Two examples of application of the method to thermodynamic data series are reported.

Towards a Multidimensional Visualization of Environmental Soft Sensors Predictions

Soft sensors have been successfully applied to simulate physical and chemical measurements in specific locations of a monument. They allowed monitoring the quality conditions of the monument surface for long periods of time. This is a not invasive process that provides a huge set of multidimensional data. They have been analyzed by the Cultural Heritage experts to find physical or chemical critical condition which could generate some degradation process. Here we propose several multidimensional visualization techniques to represent the predictions of environmental parameters, given by several soft sensors, in a comprehensive and compact way. Visualization tools using both shape variation (glyph) and color are developed to realize an homogeneous communication paradigm. Moreover, each tool uses a 3D navigable model of the monument as visual support.

WAVELETS AND SPLINES FOR VERTICAL SCRATCH REMOVAL IN OLD MOVIE SEQUENCES

In old movies, scratches are common damages that mostly result from a mechanical stress during the projection. A method for removing vertical scratches is proposed and suitable to be automatically applied to sequences of images. The method uses a wavelet decomposition of the original digital image, I, in order to separate the high frequency components and to elaborate corrupted data in the regular matrix, A, and in the vertical details matrix, V, only. For A, approximating functions are constructed in suitable spline spaces, which depend on the morphological quality of the image near the scratch. Monochromatic old images and images with simulated scratches have been considered to validate the method.

MULTILEVEL RESTORATION METHOD FOR LINE DEFECTS IN DIGITIZED IMAGES

Line defects often affect the quality of digitized images. Examples are vertical scratches in images from old movies and photos or systematic disturbances due to the preparation and acquisition of images by means of a scanner in several application fields such as biology or nanotechnology. Here an automatic method is introduced, which depends only on the parameters of the defect, but does not require a defect model. It is based on two mathematical tools: the stationary wavelet decomposition for a good treatment of the high frequencies of the signal, weighted median filters to correct the data at different levels of decomposition using iocal information. The method is validated as applied to virtually remove defects of various origins in images from old movies, photos and images from microscope.