Image Classification Using Biomimetic Pattern Recognition with Convolutional Neural Networks Features

Table 1

Comparison of traditional pattern recognition and BPR.

	Traditional pattern recognition	Biomimetic pattern recognition
Starting point	Optimal classification of different classes	Recognition of samples one by one class
Theoretical basis	All available information is included in the training set	Continuity of one sample class in feature space
Math tool	Statistics	Topology
Analyze methods	Theoretical derivation of algebra and equations (logical thinking)	Descriptive geometry of high-dimensional space (imagery thinking)
Recognition method	Division	Coverage of complex geometry in high-dimensional space
Realization approach	SVM and traditional neural networks	Multiweight high-order neural networks