Low Power Approximate 3:2 and 4:2 Compressor Base Designed Gaussian Filter for Image Processing Applications

This paper presents an optimized Gaussian filter architecture implemented with custom -designed approximate 3:2 and 4:2 compressors to improve power efficiency in image processing systems.The design is evaluated using Verilog HDL and simulated on Xilinx Vivado.Comparative analysis with conventional exact compressor designs shows significant improvements in power consumption and delay while maintaining acceptable image quality.MATLAB -based error analysis using PSNR, MRED, NMED metrics confirms the suitability of the proposed architecture for error -resilient applications such as IoT cameras and real -time edge detection systems.The results demonstrate that the proposed design achieves up to 33.33% power savings and 7.97 % delay reduction compared to exact methods, making it strong for low power embedded vision systems.