Introduction to Seminar Topic on Artificial Intelligence Pattern Recognition Using Neural Networks:

Innovative technology are the ones that takes the world forward, computer was one such innovation, it made human work pacier and error free. It is a common fact that electronic devices lack the ability to think and act according to a particular situation; this is because they don’t have any intelligence of their own and once it is loaded by code feeds it will behave intelligently and this method of feeding knowledge to an electronic system is called artificial intelligence. One such type of artificial intelligence is the pattern recognition of the images with the help of numerous neural networks. It is a simple technique where the system adapts to the object send information.

A non-patterned sample of a neural network is hard to be trained as their population will be too large. A basic generalization can be made up from this pattern which helps them to learn the concerned relations.  During pattern recognition process, the input image pyramid is subjected to sub sampling first and extracted windows are formed of 20*20 pixels. The pixels then go to a pre processing mode which includes correcting the lighting and histogram equalization.

 The neural networks then take these pre processed image pixels and assign it as network input. It passes through respective fields and come out as hidden unit. The final unit is given as output. The discriminant s would contain the basic structure of pattern recognition. The whole process of involving neural networks is noise tolerant and this results it in extensive usage in pattern recognition field.

Pattern recognition is also used in speech recognition by reading the lip movement and acoustic patterns. This technology is also used to detect facial expressions and also used in security systems in banks and companies. Pattern recognition technology using neural networks is definitely one to look out for the future. 

Download  Seminar Report on Artificial Intelligence Pattern Recognition Using Neural Networks  .