ABSTRACT
This paper describes a real-time video image processing system, suitable for image analysis of stationary and moving images. It consists of a high-quality microscope, a general-purpose personal computer, a commercially available image-processing hardware module plugged into the computer bus, a b/w TV-camera, video monitors and a software package. The structure and the capability of this system are explained. The software is menu-driven and performs real-time image enhancements, real-time mathematical and morphological filters, image segmentation and labelling, real-time identification of moving objects, and real-time analysis of their movements. The program is available in listing form.
Subject(s)
Image Processing, Computer-Assisted , Microscopy/methods , Video Recording/methods , Algorithms , Computer Systems , Forecasting , Image Processing, Computer-Assisted/trends , Microcomputers , Programming Languages , Software , Software DesignABSTRACT
The induction period of Oxytricha bifaria, namely the time lag elapsing between the beginning of the inducing (= overcrowded) conditions and the formation of the first giant was studied. It was found that this period consists of at least three successive steps: (a) the "early activation window", roughly corresponding to the first third of the induction period, during which the cell to cell contacts triggering the differentiation must occur; (b) the "late activation window", which corresponds to the central part of the induction period and during which a specific protein synthesis occurs. These two steps end in the "activation" of Oxytricha, which is now ready for the third phase, (c) "prédation" which occurs during the last third of the induction period, when the activated, potential giants actually engulf their preys.
ABSTRACT
A program for a detailed and fast feature extraction procedure on objects from a digitized microscope image is presented. The single-pass algorithm labels the pixels of the image that are above a selected threshold, and connects those which belong to the same object. A data structure addressable as an associative memory is created, thus avoiding the use of any recursive procedures. A detailed description of the procedure, which extracts features such as integrated optical density, area and baricenter coordinates, is presented.