Speaker
Description
Enhancing object recognition by optimizing sharpness
Deeva Olga(a), Inna Kolesnikova(b), Oksana Streltsova(b), Marko Ćosić(c)
(a) National Research Nuclear University MEPhI
(b) Joint Institute for Nuclear Research, Laboratory of Radiation Biology, Dubna, Str. Joliot-Curie 6, Russia
(c) Institute of Nuclear Sciences Vinča, National Institute of Republic of Serbia, University of Belgrade, Belgrade, Serbia
The task of recognizing objects in an image is relevant in many spheres of human activity. Since the image can be distorted by noise, blurred or be poorly-lit, it is necessary to use preprocessing for high-quality recognition of objects. In our software the sharpness, which was defined as a maximazed gradient, of the monochromatic image was found depending on the percentage of rgb channels when added to a gray image. According to the data, 3d graphs of the sharpness dependence on the channels were constructed. The obtained result can be used to sharpen any images. The initial idea of this algorithm was to sharpen the images of hippocampus (that consist of neuron cells, vessels and sometimes artefacts), so that these changed images could be used as a data for further studies. Such preprocessing could be a great part of a future contemporary tool for pathomorphologists as well as for scientists in this branch.
