Title :
A classifier system with low sensitivity to pattern shifted position
Author :
Nino-de-Rivera, L. ; Perez-Meana, Hector ; Sanchez-Sinencio, Edgar
Author_Institution :
Dept. of Electr. Eng., Univ. Autonoma Metropolitana, Mexico City, Mexico
Abstract :
This paper presents a proposal for dedicated hardware classifier which implementation seems to be feasible for analog VLSI processing. Analog distance and correlation metrics are adopted and its architecture to be implemented in VLSI is discussed. Analog distance reports face identification for face rotation until 30 degrees. Correlation lets identification for shifted position human face in a x-y plane. Simple structures are presented for analog distance and correlation. Analog distance can be used as an algorithm in a PC to classify human faces from a video camera
Keywords :
VLSI; analogue integrated circuits; analogue processing circuits; correlation methods; face recognition; pattern classification; PC; algorithm; analog VLSI processing; analog correlation metric; analog distance metric; dedicated hardware classifier; human face identification; pattern shifted position; sensitivity; video camera; Cameras; Difference equations; Face; Humans; Neural network hardware; Neural networks; Proposals; Testing; Very large scale integration; Voltage;
Conference_Titel :
Circuits and Systems, 1996., IEEE 39th Midwest symposium on
Conference_Location :
Ames, IA
Print_ISBN :
0-7803-3636-4
DOI :
10.1109/MWSCAS.1996.588047
