A High Sensitive and Fast Motion Estimation for One Bit Transformation Using SSD

In this paper, architectures implementing fixed block size (FBS) and variable block size (VBS) for One bit transform (1-BT) using SSD (Sum of Squared Differences) to find motion between frames have been proposed. In general Motion estimation (ME) requires a huge amount of computation, and hence consumes the largest amount of power. The technique used to reduce the usage of power is 1-BT based ME algorithms. They have low computational complexity. In this less On-chip memory used than the previous 1BT based ME hardware by using a data reuse scheme and memory organization. Here the Sum of Absolute Difference (SAD) matching method is followed with the application of Diamond Search (DS). The 1-BT based ME is usually performed by applying full search (FS) algorithm. The proposed architecture reduce the power by performing the application of toroidal path in the pipelining approach on processing element with Sum Of Squared differences technique(SSD). The simulation results reveal that the application of DS on 1-BT based ME using SSD technique can significantly reduce the computational complexity and the power which is observed in SAD based 1-BT ME.