F. Russo, A. Lazzari
AUTOMATIC REDUCTION OF NOISE IN DATA ACQUIRED BY DIGITAL CAMERAS

A technique for automatic noise correction in color images acquired by digital cameras is presented. The proposed method deals with the R, G and B components of the image and adopts a nonlinear approach in order to perform a very accurate restoration of the data. The optimal parameters that define the key filtering action are obtained by an automatic procedure that independently operates on each channel. Experimental results show that the method is very effective and well suited to correct the typical noise increase that occurs when high sensitivity is chosen during image acquisition.

P. A. Traverso, G. Pasini, D. Mirri, F. Filicori, G. Iuculano
EXPERIMENTAL VALIDATION OF A GENERAL-PURPOSE NON-LINEAR DYNAMIC MODEL FOR S/H-ADC DEVICES

Experimental results are provided in this paper in order to validate the Discrete-Time Convolution approach, recently proposed for the modelling of both static and dynamic non-linearities in S/H-ADC devices. The comparison between actual device response and model predictions is provided, showing a very good agreement under operation conditions which are not suitably described by a conventional model, based on the characterisation of the static non-linearity only. A brief discussion about the possibility of embedding into the model almost any recent empirical technique for the characterisation of static sources of uncertainty within the device is proposed as well.

D. Dallet, Ch. Rebai, V. L. Boun, D. Elfersi, A. Merino, Ph. Marchegay
Adaptive Filtering For ADC Spectral Analysis

In this paper, we present an adaptive filter structure for Analog to Digital Converters (ADC) spectral domain analysis. A digital filter bank is commonly used to decompose an ADC output signal into its main spectral components. However, this method requires a priori accurate knowledge of the output characteristics. It implies that a slightly deviation from the calculated resonant frequency results in a drastic decrease in the parameters estimation precision. In order to solve this problem, we propose a promising adaptive structure for a Built-In-Self-Test (BIST) approach for ADC and Mixed-Signal Integrated Circuits (IC).

F. Attivissimo, N. Giaquinto, M. Savino
EVALUATING MEASUREMENT UNCERTAINTY IN A/D CONVERTERS WITH AND WITHOUT DITHER

The paper gives formulae for evaluating the uncertainty of measurements (both “direct” and “indirect”, i.e. functions of measurements) performed through an ADC-based device (like an oscilloscope, a plug-in DAQ board, etc.). The way of modeling the errors and writing the specifications is discussed. The mathematically complex but nowadays common technique of dithering is also analyzed. The results are intended as a contribution towards a standard way of writing and using ADC specifications.

Raul Carneiro Martins, António Cruz Serra
Stochastic Approach for Memoryless Nonlinearity Measurements

In this paper we will present a simple and cost effective, yet accurate, setup for measuring the amplitude distribution of a signal and its probability density function (pdf ). From these, it will be shown how to measure and model linear and memoryless nonlinear errors using both a power and an orthogonal series representation of the distorted signal, and how to use this knowledge to improve the resolution of an ADC characterization. Finally, we will give an usage application, with experimental results, in which the nonlinearities of a low grade stimulus are ‘subtracted’ in the histogram method for the characterization of ADCs.