ИСТИНА

Low signal-to-noise ratio and upredictability of timing and location of spontaneous calcium signalling events in astrocytic networks challenges their reliable detection and reconstruction from two-photon imaging data. We suggest a modified multiscale vision model, an object detection framework based on thresholding of wavelet coefficients and hierarchical trees of significant coefficients followed by non-linear iterative partial object reconstruction, for analysis of two-photon calcium imaging data. We extend the framework by the mixed multiscale median/starlet transform for image decomposition and iterative reconstruction of the detected objects. Comparison with several popular state of the art image denoising methods (total variation denoising, bilateral filtering and SURE-LET denoising) shows that for the tested patterns the multiscale vision model has similar performance in the denoising aspect, but provides better segmenation of the image into meaningful objects, whereas other methods need to be combined with dedicated thresholding and segmentation utilities.