An Efficient Medical Assistive Diagnostic Algorithm for Visualisation of Structural and Tissue Details in CT and MRI Fusion

• Published in: Cognitive computation Vol. 13; no. 6; pp. 1471 - 1483
• Main Authors: Goyal, Bhawna, Dogra, Ayush, Khoond, Rahul, Al-Turjman, Fadi
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.11.2021; Springer Nature B.V
• Subjects: Algorithms; Artificial Intelligence; Computation by Abstract Devices; Computational Biology/Bioinformatics; Computer Science; Computer vision; Decomposition; Image contrast; Image enhancement; Image filters; Image segmentation; Localization; Magnetic resonance imaging; Medical diagnosis; Medical imaging; Methods; Sensors; Striations; Tomography; Visual perception; Wavelet transforms; CT; MRI; Imaging; Medical; Diagnosis; Tumour; Fusion
• ISSN: 1866-9956; 1866-9964
• DOI: 10.1007/s12559-021-09958-y

Clinicians often have to switch amongst radiographic scans in order to trace out patterns in various tissue striations. The conglomerated view of structural and anatomical view in medical scans can facilitate the physicians to execute precise diagnosis, intraoperative guidance, and planning preoperative procedures. Due to inherent physical limitations, source images have prevalence of noise and ambient light. This results in lower contrast and limited visual perception of striations and tissues in fused radiographic images. This paper proposes a concatenated filtering image fusion approach employing space segmentation and non-prior-based contrast enhancement. The latent row rank theory approach implements sub-space segmentation addressing the issue of noise removal, and the non-local-prior-based enhancement removes the ambient light from source images fortifying edge details and information. This complex fusion framework is designed in non-sub-sampled contourlet transform which exhibits computational efficiency. The final fused image obtained using local Laplacian energy fusion rule results in improved localisation of structural and anatomical details of brain tissue and outperforms high-performing fusion methods in literature both objectively with high fusion rate along with better quality visual results.