Subspace-Based Face Recognition in Analog VLSI

We describe an analog-VLSI neural network for face recognition based on subspace methods. The system uses a dimensionality-reduction network whose coefficients can be either programmed or learned on-chip to perform PCA, or programmed to perform LDA. A second network with userprogrammed coefficients performs classification with Manhattan distances. The system uses on-chip compensation techniques to reduce the effects of device mismatch. Using the ORL database with 12x12-pixel images, our circuit achieves up to 85% classification performance (98% of an equivalent software implementation).

• Type: Conference
• Year: 2007
• Where: NIPS
• Authors: Gonzalo Carvajal, Waldo Valenzuela, Miguel Figueroa
• Keywords: Analog-VLSI Neural Network, Information Technology, NIPS 2007, On-chip Compensation Techniques, Userprogrammed Coefficients Performs