Design of Efficient Median Filter in Quantum-dot Cellular Automata with Energy Dissipation

Abstract

The Quantum-dot Cellular Automata (QCA) technology is introduced to overcome the limitations such as material, size, speed, feature, scalability, switching frequency, and power consumption over traditionally well-known CMOS technology used for Very Large-Scale Integration (VLSI) circuits design. The QCA represents the digital information by the polarization of electrons. Elastic relaxation produces quantum dots, which have direct interaction with the optical and electrical features of quantum dot-based systems. The Coulomb interaction and quantum phenomena are used in QCA technology [1].

Generally, quantum dot arrays are used to implement distinct Boolean logic functions in QCA technology. Combinational circuits have been implemented by arranging quantum cells with appropriate clock delays. The power consumption of QCA based circuits are significantly affected by clock signals. The power consumption will be quite low in practice when utilizing a proper clock, although it is still data dependant. Bennett clocking system can be used for various inputs to remove data dependency of power traces [2]. These benefits can be extremely advantageous in many real-time image processing applications. The presence of impulse noise degrades the qualitative performance of Digital Image Processing (DIP) techniques. Images are corrupted by impulse noise which is caused by malfunctioning either camera sensors or lack of memory in hardware. The corrupted images have been recovered by using a median filter.

The application of QCA technology in digital image processing has been reported by Bandan et al. [3] by proposing a median finder architecture for removing impulse noise in corrupted images. A one-hot encoder and majority gates have been used for the implementation of median finder architecture. In the architecture proposed by Bandan et al. [3], every number has been converted in the form of one-hot encoding. The majority number has been selected in each column by using the majority gate. The selected number has been considered as a median value. This architecture needs an external encoder circuit for converting numbers from binary form to one-hot encoding form. Therefore, the architecture occupies more area which is the main drawback of using this architecture for finding the median value.

This paper mainly demonstrates the implementation and simulation results of Median Filter (MF) architecture using Quantum-dot Cellular Automata (QCA) technology. The MF plays an important role in DIP for removing impulse noise. The proposed QCA based MF have single-layered with less cell count and low latency. Further, the MF is designed by using Compare and Selective Module (CSM). The CSM is taking the decision based on the comparison of the two inputs given. The QCA Designer – E simulation tool has been used to design & verify all the proposed architectures. The energy dissipation has been simulated using a coherent vector engine setup. The implementation results of the proposed 1-bit & 2-bit CSM architectures occupy the area of 0.17 & 0.52 µm² and use 118 & 380 QCA cells, respectively. The proposed CSM is further extended to a larger bit size. The 1-bit & 2-bit CSM architectures have the total energy dissipation is 4.13 × 10^-2 & 1.46 × 10^-1 eV, and the average energy dissipation is 3.76 × 10^-3 & 3.76 × 10^-2 eV, respectively. The total & average energy dissipation per cycle of proposed MF is 41.72 × 10^-1 & 38.26 × 10^-2 eV, respectively.