Implementation of Realtime Image Fusion for Biomedical Applications Using ICA And Discrete Wavelet Transform

Abstract

Image fusion is an extensively used technique in various areas like computer visualization, enhanced diagnostic imaging, radio therapy, automatic object recognition, image analysis, and remote sensing. The main aim of image fusion is to combine several input images into one image containing more information than the individual images. This type of image fusion results in a new image that is easier for computers and humans to see, making it possible for additional image processing operations like object detection, segmentation, and feature extraction. This paper examines the potential application of customized wavelet transform for image fusion. This paper is implemented for biomedical applications by using wavelet-based pixel-level image fusion technique and Independent Component Analysis (ICA) for de-noising. For realization of this, Initially, MRI and CT pictures are extracted using the 2D discrete wavelet transform and then ICA is used to obtain ICA bases and fused on the sub band pictures. The suggested methodology is implemented in FPGA Spartan and Virtex devices since FPGA provides a good platform for image processing in real-time and finally the performance has been compared with the existing methodologies.