Optical Recognition Objectives

Technology

Our software for pattern recognition is based on the absolutely novel GABITUS technology. Methods of GABITUS arise from the artificial competitive neural network model, developed under the joint research with Samara Technical University. This neural network allows interpreting image as the whole integral pattern rather than a set of sample elements by the geometrical and semantic standpoint.

Due to the developed approach, GABITUS enables to obtain results very close to the associative human thinking, while they are not rely on formal statistic calculations.

The recognition follows to the search for the prototype that match the offered pattern among all samples memorized before. When the pattern is being memorized, our neural network builds the associative connections as the mutual similarity levels, that leads to the complex nonlinear dynamics in this neural network model. This seriously improves the recognition results.

The GABITUS technology is the powered engine of the unique information system CRIMNET made up by ORO LLC. for person identification purposes. It serves for the processing of huge information amounts containing photos and composite pictures in cases of searching faces, which look like the offered picture. The effectiveness of our technology opens the possibility of its installation on desktop PCs while the fundamental level of recognition and identification problem solving is achieved.

The developed technology uses the outstanding fast recognition technique and may be implemented in the wide range of tasks of complicated information processing. It can be realized in the software for PCs as well as multiprocessor servers.

If you need the information system that thinks rather than just performs the loaded program, you inevitably choose the GABITUS.