| Title |
Cosine colorization for enhancement of Magnetic Resonance Angiography (MRA) |
| Authors |
Odusami, Modupe Olufunke ; Damaševičius, Robertas ; Abayomi Alli, Olusola Oluwakemi ; Maskeliūnas, Rytis |
| DOI |
10.15439/2026F8443 |
| eISBN |
9788397708846 |
| Full Text |
|
| Is Part of |
Proceedings of the 21st Conference on Computer Science and Intelligence Systems (FedCSIS), 23-26 August 2026, Riga, Latvia.. Warsaw : Polish Information Processing Society, 2026. p. 81-88.. eISBN 9788397708846 |
| Keywords [eng] |
Cosine colorization ; enhancement ; Magnetic Resonance Angiography ; vessel-restricted mapping |
| Abstract [eng] |
Non-invasive imaging modalities play a crucial role in healthcare. The use of Magnetic Resonance Angiography (MRA) has enhanced access to blood vessel information. MRA grayscale preserves anatomical fidelity but offers limited perceptual separation. This study investigates the use of cosine colorization to enhance feature discrimination in MRA. The cosine colorization utilizes the Frangi vesselness filter combined with statistical thresholding to isolate vascular regions. Quantitative and statistical evaluations were conducted on 20 MRA volumes, comparing grayscale baselines with full-volume and vessel-restricted cosine mappings. While full-volume colorization introduces structural distortion (SSIM ≈ 0.85 and NMSE ≈ 0.018), the vessel-restricted approach maintains near-perfect anatomical fidelity (SSIM ≈ 1.0 and NMSE ≈ 0). Statistical analysis confirms significant improvements in perceptual separability (p < 10−5) without degradation of structural metrics for the vessel-restricted cosine approach. The results show that the vessel-restricted cosine approach provides an interpretable enhancement of vascular structures. |
| Published |
Warsaw : Polish Information Processing Society, 2026 |
| Type |
Conference paper |
| Language |
English |
| Publication date |
2026 |
