Online processing of microscopic images

In this paper we present the concept and prototype for online automatic processing of microscopic images. The idea behind the system is based on the needs of different scientific and professional areas for automation of image processing. There are many tasks in these areas that can take advantage of computer vision (CV) and machine learning (ML) techniques for image segmentation and classification. Advances in ML area of deep neural networks enable us to use the same models for many different recognition tasks. The problem is that image processing usually requires particular skills in CV and ML for successfully building these models. With this system, all models would be supervised by independent ML experts while other users could set up systems through simplistic wizards. Another problem is obtaining and preparing large quantities of data for training and testing CV and ML models. Manual sampling and labelling of large quantities of images is necessary for building good automatic models. This problem can be mitigated by online collaboration between interested parties through sharing image databases.