论文信息 - Synthesis of a recurrent double-layer transistor network for early-vision tasks

Synthesis of a recurrent double-layer transistor network for early-vision tasks

In this paper we present a double-layer transistor network, useful for early-vision embedded systems, able to convolve a sensorial input with a Gabor kernel. The network, whose main advantages are low power consumption, compactness of physical implementation and full programmability of the convolution kernel, has been simulated and successfully compared with theoretical expectations.

Luigi Raffo | Massimo Barbaro | A. Nazzaro

[1] Kwabena Boahen,et al. Translinear circuits in subthreshold MOS , 1996 .

[2] Tomaso Poggio,et al. Computational vision and regularization theory , 1985, Nature.

[3] Tetsuya Yagi,et al. Image processing regularization filters on layered architecture , 1993, Neural Networks.

[4] J. L. White,et al. An active resistor network for Gaussian filtering of images , 1991 .

[5] Tomaso Poggio,et al. Computational vision and regularization theory , 1985, Nature.

[6] G. M. Bo,et al. Visual Perception Microsystems Based on Distributed Analog Vlsi Processing , 1997 .

[7] Eric A. Vittoz. Pseudo-Resistive Networks and their Applications to Analog Collective Computation , 1997, ICANN.

[8] Daniele D. Caviglia,et al. Anisotropic active resistor meshes for implementing image processing operators , 1993 .

[9] Luigi Raffo. Analysis and synthesis of resistive networks for distributed visual elaborations , 1996 .

[10] Luigi Raffo,et al. Analog VLSI circuits as physical structures for perception in early visual tasks , 1998, IEEE Trans. Neural Networks.

[11] Andreas G. Andreou,et al. A Contrast Sensitive Silicon Retina with Reciprocal Synapses , 1991, NIPS.

[12] J. L. White,et al. Active resistor networks as 2D sampled data filters , 1992 .

[13] G. M. Bob,et al. A Current-mode Computational Engine for Stereo Disparity and Early Vision Tasks , 1997 .

[14] Eric A. Vittoz,et al. Analog VLSI signal processing: Why, where, and how? , 1994 .