Scholarlink

<p>Multimodal medical image fusion integrates complementary information from different imaging modalities, such as MRI and CT, to improvediagnostic accuracy while reducing redundant information. However, conventional fusion approaches often require large annotated datasets,suffer from loss of edge and texture details, and are computationally intensive, limiting their applicability in real-time clinical settings. In this study,a hybrid fusion framework based on Continuous Wavelet Transform (CWT), Dual-Tree Complex Wavelet Transform (DTCWT), and PhaseCongruency (PC), termed CWT-DTCWT-PC, is proposed. Pre-registered MRI and CT images from the Brain Tumor Multimodal Image Dataset aredecomposed using multiscale transforms. CWT enables multi-scale feature extraction with reduced shift variance, while DTCWT capturesdirectional and phase information with near shift-invariant properties. Phase congruency is employed to select structurally signiicant coeficients,ensuring preservation of salient edges and textures during fusion. Feature-preserving fusion rules are then applied to generate a single high-quality fused image containing critical diagnostic information from both modalities. Experimental results demonstrate that the proposed methodachieves superior performance compared to recent deep learning-based fusion models, attaining a PSNR of 48 dB and improved structural detailpreservation. The proposed framework offers a computationally eficient and robust solution for multimodal medical image fusion.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

CWT and phase congruency moments based multimodal image fusion

  • 期刊论文

  • 发表日期：

    2026年3月30日

查看全文
  • Arathi T,
  • Naveen S,
  • Rahul C,
  • Aneesh M H

摘要

Multimodal medical image fusion integrates complementary information from different imaging modalities, such as MRI and CT, to improvediagnostic accuracy while reducing redundant information. However, conventional fusion approaches often require large annotated datasets,suffer from loss of edge and texture details, and are computationally intensive, limiting their applicability in real-time clinical settings. In this study,a hybrid fusion framework based on Continuous Wavelet Transform (CWT), Dual-Tree Complex Wavelet Transform (DTCWT), and PhaseCongruency (PC), termed CWT-DTCWT-PC, is proposed. Pre-registered MRI and CT images from the Brain Tumor Multimodal Image Dataset aredecomposed using multiscale transforms. CWT enables multi-scale feature extraction with reduced shift variance, while DTCWT capturesdirectional and phase information with near shift-invariant properties. Phase congruency is employed to select structurally signiicant coeficients,ensuring preservation of salient edges and textures during fusion. Feature-preserving fusion rules are then applied to generate a single high-quality fused image containing critical diagnostic information from both modalities. Experimental results demonstrate that the proposed methodachieves superior performance compared to recent deep learning-based fusion models, attaining a PSNR of 48 dB and improved structural detailpreservation. The proposed framework offers a computationally eficient and robust solution for multimodal medical image fusion.