Real-time cloud computing for web-based searching system of pattern recognition

Abstract

For cross-platform real-time systems, cloud computing technology is an innovative application of pattern recognition method. This study is the use of associative memory of the way to do the work of pattern recognition; this system is real-time client-server type network pattern recognition system. Remote user can operate through the browser to draw the shape or character of industrial components, and recognition system to the database through Internet search. Cloud storage server contains a database of pattern samples. In the training period, the user can specify any of the pattern to what in real-time. Patterns are recorded in the cloud server database. In the recall period, an innovative database matching methods have been proposed. This method can effectively solve the problem of RNN a false state of the database than on the technology to overcome the problem of capacity constraints RNN. In this new approach, CWBPR system partition database in the cloud server, a pattern record set, and then figure out they were separate sections for each value of W and θ. CWBPR system to deal with each of the last segment of the pattern recognition work. Pattern recognition technology for the network, the paper has two simulation experiments are clearly discussed. The first experiment identified a number of characters; the second experiment is the pattern recognition of industrial components. Finally, the paper also put forward innovative pattern recognition method to the traditional text input search method comparison.