General automatic detection systems that are used to check acceptability and the quality of agricultural and industrial products are built on the basis of sensory assessment standards by professional inspectors. However, assessment standards of products vary according to the ability or fatigue levels of such inspectors, and it is difficult to build systems that offer constant judgment standards.
The neural network (NN) of the self-organizing type we devised has the capability to automatically extract statistical characteristics of data given and stochastically classify the data into several groups.
By using the NN, without advance research regarding product quality standard, it is expected that development of new automatic product assessment systems which enables check of acceptability and quality in objective and universal manners will be possible.
Created ： 2008/04/01 10:13 Modified ： 2015/04/06 15:28