signal and measured data because of the instrumental inexelence ??measured datas differs from true signal. The presence of the noise is one of the main factors that make the restoration of true signal difficult how the symmetry of the equation that describe the procces of data formation can be used to eliminate the noise. WHERE THIS PROBLEM OCCUR different kind of telescopy image proccessing, electrophoresis, HOW IT MAY BE MATHEMATICALY FORMULATED Finite resolution -> discrete data sets Values of the true signal emmited from points of a certain object are stored in one set and observed datas in the second set of the same size. because of measurement procces the true signal from each single point of the object is spreeded to different points in the data set. Mathematicaly this bluring can be described by the equation B=K*O in which O is diagonal matrix with entries equal to the values of the signal K is the matrix that describe spreed, from each point of the object can reach different points in the set of datas Without the noise the set O and D of points object and observed datas are represented by sets of points. signal from each point of the object can reach different points in the set of datas true value of the signal in each point of the signal from each point of the object can reach different points in the set of datas how the spred appears true signal array Spreed array Spreeded signal Array noise array how the noise appears SPRED AND NOISE discussion -> question of solvability Spreed array can be calibrated, tested, and is known with some accuracy, ussualy gaussian core and and so on Known about noise unknown it is ussualy only assumed to posses certain mean value and distribution. disadvantages of existing methods filtering is a compromise beetwen removing the noise and damaging true signal. The fitting yields only aproximation, which may be far from the signal (and may contain undesirable artifacts) SYMMETRY - additional information about the signal