Convolutional spiking neural network model for robust face detection

Abstract

We propose a convolutional spiking neural network (CSNN) model with population coding for robust face detection. Basic structure of the network includes hierarchically alternating layers for feature detection and feature pooling. The proposed model implements hierarchical template matchmg by temporal integration of structured pulse packet. The packet signal represents some intermediate or complex visual feature (e.g., a pair of line segments, comers, eye, nose, etc.) that constitutes a face model. The output pulse of a feature pooling neuron represents some local feature (e.g., line segments). Introducing a population coding scheme in the CSNN architecture, we show how the biologically inspired model attains invariance to changes in size and position of face and ensures the efficiency of face detection.

Selection and Efficient Use of Local Features for Face and Facial Expression Recognition in a Cortical Architecture

“…CoNN (LeCun and Bengio, 1995) as well as Neocognitrons (Fukushima, 1980) have been used for face detection (Matsugu et al, 2002;Osadchy et al, 2004) and recognition (Lawrence et al, 1995). Proposed architecture in Figure 1 comes with the property of robustness in object recognition such as translation and deformation invariance as in well-known neocognitrons, which also have similar architecture.…”

“…First, it has only FD modules in the bottom and top layers. The intermediate features detected in FD2 constitute a set of figural alphabets (Matsugu et al, 2002;Matsugu & Cardon, 2004). Local features in FD1 are used as bases of figural alphabets, which are used for eye or mouth detection.…”

“…The training proceeds as follows. As in (Matsugu et al, 2002, training of the MCoNN is performed module by module using fragment images as positive data extracted from publicly available database (e.g., Softpia Japan) of more than 100 persons. Other irrelevant fragment images extracted from background images are used as negative samples.…”

“…In this chapter, inspired by cortical processing, we will address the problem of efficient selection and economical use of visual features for face recognition (FR) as well as facial expression recognition (FER). We demonstrate that by training our previously proposed (Matsugu et al, 2002) hierarchical neural network architecture (modified convolutional neural networks: MCoNN) for face detection (FD), higher order visual function such as FR and FER can be organized for shared use of such local features. The MCoNN is different from those previously proposed networks in that training is done layer by layer for intermediate as well as global features with resulting receptive field size of neurons being larger for higher layers.…”

A Convolutional Neural Network VLSI for Image Recognition Using Merged/Mixed Analog-Digital Architecture

“…Convolutional neural networks with a hierarchical structure, which imitate the vision nerve system in the brain, have such functions [1][2][3].…”