Fowlkes-Mallows Correlated Cohen Kappa Coefficient Block Matching based Multi-Layer Perceptron Classification for Motion Estimation in VLSI
M Sunitha, Mary Valantina G · International Journal of Electrical and Electronics Engineering · 2023
Motion Estimation (ME) constitutes an essential process in video coding with lesser processing time. Many methods were designed for efficient motion estimation in the VLSI architecture. However, the consumption of power was not reduced through existing techniques. Fowlkes–Mallows Correlated Cohen Kappa Coefficient Block Matching-based Multi-Layer Perceptron Classifier (FMCCKCBM-MLPC) Model is introduced to handle such limitations. FMCCKCBM-MLPC Model is used for increasing the motion estimation of video series in the VLSI architecture circuits. Multi-Layer Percepton is used for examining the feature and performing classification with the multiple layers. Input is sent to the hidden layer 1. In hidden layer 1, the segmentation process is performed by Fowlkes–Mallows Regularized Principal Component Regression to reduce the time loss. In hidden layer 2, the block matching is carried out by Cohen Kappa Coefficient with the segmented blocks. With Cohen Kappa Coefficient, the repeated blocks are predicted with lesser loss in the video series. This helps to enhance motion estimation in the VLSI architecture. FMCCKCBM-MLPC Model is computed in terms of power and time consumption. The simulation result of the FMCCKCBM-MLPC Model minimizes the area, PSNR, delay and power consumption of motion estimation in the video series with existing techniques.