Ultra low energy analog image processing using spin based neurons

Abstract

In this work we present an ultra low energy, onsensor image processing architecture, based on cellular network of spin based neurons. The neuron constitutes of a lateral spin valve (LSV) with multiple input magnets, connected to an output magnet, using metal channels. The low resistance, magneto-metallic neurons operate at a small terminal voltage of ~20mV, while performing analog computation upon photo sensor inputs. The static current-flow across the device terminals is limited to small periods, corresponding to magnet switching time, and, is determined by a low duty-cycle systemclock. Thus, the energy-cost of analog-mode processing, inevitable in most image sensing applications, is reduced and made comparable to that of dynamic and leakage power consumption in peripheral CMOS units. Performance of the proposed architecture for some common image sensing and processing applications like, feature extraction, halftone compression and digitization, have been obtained through physics based device simulation framework, coupled with SPICE. Results indicate that the proposed design scheme can achieve more than two orders of magnitude reduction in computation energy, as compared to the state of art CMOS designs, that are based on conventional mixed-signal image acquisition and processing schemes. knowledge, this is the first work where application of nano magnets (in ) in analog signal processing has been proposed.

Soft computing-based modeling of flotation processes – A review

“…Cellular neural network can be regarded as a fusion of an artificial neural network and cellular automata (Sharad et al, 2012). There are reports on the application of such networks in modeling the parameters of coal flotation concentration, due to their ability to quickly process flotation froth images (Zimmermann and Jeanmeure, 1996;Jeanmeure and Zimmermann, 1998).…”

Memristive Threshold Logic Circuit Design of Fast Moving Object Detection

“…Although objects are detected in all the frames, those that are difficult to recognise due to overlap with other objects in the same frame are considered to be not detected to make it realisable in practical object tracking tasks. There exists low-power programmable circuits mimicking neural behaviour [25], [26] that can be used for object detection applications. However, any objective comparison between the proposed cell and those of the like in neuron-synapse in [25] is difficult with different technologies without reporting the power and area specific at the cell level.…”

Exploring Boolean and non-Boolean computing with spin torque devices

“…Ultra low voltage, current-mode operation of magneto-metallic devices like LSV's and domain-wall magnets (DWM) can be used to realize analog summation/integration and thresholding operations with the help of appropriate circuits [17][18][19][20][21][22], [29][30]. Such devicecircuit co-design can lead to ultra-low power non-Boolean computation circuits.…”

“…Design Example: The circuit concept described above can be employed in the design of different classes of neuromorphic architectures. Using this technique, we presented the design of an image processing architecture based on cellular neural network (CNN) in [30] ( fig. 6).…”