One of the problems dealt with is the pattern recognition in blood cell microphotographs. The blood cell count for various types of cells is extremely important for the diagnostics of many diseases. Currently available methods are either rather complicated, or leave undetected some of the rarely appearing (but often very dangerous) cell types. The wavelet method as applied by the group helps determine all the cells of various morphology and count them automatically. A special computer program has been developed and tested which focuses a microscope on the blood sample, classifies the blood cells by their different shapes, registers and stores these counts in the computer. The complete blood analysis is done quickly and automatically. This method has been developed in collaboration with Moscow Medical Academy. Encouraging results have been obtained in the studies of skin, teeth and eyes by the femtosecond methods of the coherent optical tomography. MWA has been successfully applied to pattern recognition of various anomalies and their diagnosis. For the first time this method has permitted, for example, to determine the location and type of various local anomalies in the human skin at depths of up to 1.5 mm from its surface or unseen cavities inside teeth. Among other advantages this new method allows to avoid the drawbacks of X-ray studies on human tissue.

The following example shows how the original bad quality microphotograph of cells is wavelet-transformed to the image with sharp edges. Red color corresponds to positive wavelet coefficients, and green to the negative ones, with edges located in between.

Original cell image. Wavelet treated image.
Original cell image Wavelet treated image