VLSI Based Medical Image Fusion Using Translation Invariant Wavelet Transform with Average Fusion Rule

Abstract

The implementation of medical image fusion using VLSI design is a crucial advancement for enhancing image quality and accuracy in medical diagnostics. The global medical image processing market is projected to grow at a compound annual growth rate (CAGR) of 8.1%, reaching $47.6 billion by 2026, driven by the increasing demand for efficient diagnostic tools. Fusion techniques using wavelet transforms, particularly Discrete Wavelet Transform (DWT), have gained popularity, but they often encounter issues like poor edge preservation and loss of crucial detail, limiting their effectiveness in medical applications. In this work, we propose a novel VLSI-based implementation using Translation Invariant Wavelet Transform (TIWT) with an Average Fusion Rule. Our design incorporates a multibit adder using a 3-to-8 decoder and division using right shift, optimizing both the fusion process and overall performance. TIWT overcomes the limitations of DWT by preserving edges and eliminating artifacts that occur during image transformation. The combination of TIWT and advanced arithmetic operations allows for better detail retention in fused medical images, improving diagnostic accuracy. The proposed system achieves lower computational complexity, faster processing time, and improved fusion quality, making it ideal for real-time medical imaging applications. This approach offers three key advantages: improved edge preservation, enhanced image clarity, and reduced computational overhead. Additionally, the hardware-friendly nature of the design ensures that it was implemented efficiently in real-time systems for medical diagnostics.