TY - GEN
T1 - High-Quality Analytic Isosurface Rendering for Web-Based Cardiac Imaging
AU - Charalampidis, Vasileios Lazaros
AU - Handoko, M. Louis
AU - Ródenas-Alesinao, Eduard
AU - Asselbergs, Folkert W.
AU - Votis, Konstantinos
AU - Bizopoulos, Paschalis
AU - Triantafyllidis, Andreas
N1 - Publisher Copyright:
© 2025 IEEE.
PY - 2025
Y1 - 2025
N2 - Advancements in GPU-accelerated web-based 3D rendering have enabled high-quality, interactive medical visualization directly within standard browsers, eliminating the need for native software installations. Building on this progress, we present a fully browser-native isosurface renderer designed for volumetric medical datasets. The system supports both trilinear and tricubic interpolation, enabling sub-voxel accurate surface reconstruction through analytic root-finding techniques. To improve rendering efficiency, we extend empty-space skipping strategies to accommodate tricubic interpolation. Additionally, we introduce a novel early rejection test based on Bernstein polynomials, which efficiently identifies and excludes regions devoid of isosurface intersections-yielding up to 1.5 × speedups, especially in the computationally intensive tricubic case. While each component exists in native graphics pipelines, their integration into a browser-based framework bridges the gap between desktop-level performance and installation-free access. We validate our approach using contrast-enhanced cardiac CT volumes, achieving interactive frame rates ranging from 25 to 45 frames per second (FPS). on consumer-grade hardware. In a targeted usability study, 7 out of 8 clinicians reported that the system improved their ability to visualize patient data and communicate medical findings.
AB - Advancements in GPU-accelerated web-based 3D rendering have enabled high-quality, interactive medical visualization directly within standard browsers, eliminating the need for native software installations. Building on this progress, we present a fully browser-native isosurface renderer designed for volumetric medical datasets. The system supports both trilinear and tricubic interpolation, enabling sub-voxel accurate surface reconstruction through analytic root-finding techniques. To improve rendering efficiency, we extend empty-space skipping strategies to accommodate tricubic interpolation. Additionally, we introduce a novel early rejection test based on Bernstein polynomials, which efficiently identifies and excludes regions devoid of isosurface intersections-yielding up to 1.5 × speedups, especially in the computationally intensive tricubic case. While each component exists in native graphics pipelines, their integration into a browser-based framework bridges the gap between desktop-level performance and installation-free access. We validate our approach using contrast-enhanced cardiac CT volumes, achieving interactive frame rates ranging from 25 to 45 frames per second (FPS). on consumer-grade hardware. In a targeted usability study, 7 out of 8 clinicians reported that the system improved their ability to visualize patient data and communicate medical findings.
KW - analytic ray casting
KW - cardiac imaging
KW - empty space skipping
KW - Isosurface rendering
KW - WebGL
UR - https://www.scopus.com/pages/publications/105031167565
U2 - 10.1109/BIBE66822.2025.00019
DO - 10.1109/BIBE66822.2025.00019
M3 - Conference contribution
AN - SCOPUS:105031167565
T3 - Proceedings - 2025 IEEE 25th International Conference on Bioinformatics and Bioengineering, BIBE 2025
SP - 63
EP - 70
BT - Proceedings - 2025 IEEE 25th International Conference on Bioinformatics and Bioengineering, BIBE 2025
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 25th IEEE International Conference on Bioinformatics and Bioengineering, BIBE 2025
Y2 - 6 November 2026 through 8 November 2026
ER -