TL;DR
This paper introduces the Chinese Visible Human (CVH) datasets, a high-resolution digital representation of a Chinese male and female. The CVH project significantly improves upon previous digital human initiatives by employing advanced embalming, cryosectioning, and milling techniques to preserve anatomical detail and minimize data loss, resulting in more complete and accurate anatomical models.
What Problem Does This Paper Address?
Previous digital human projects, like the US Visible Human Project (VHP), provided valuable anatomical data but suffered from limitations. These included data loss due to cadaver segmentation, a lack of representation for diverse populations (VHP being Caucasian), and lower resolution in some areas. The CVH project aimed to address these gaps by creating detailed, high-resolution digital datasets of a typical Chinese male and female, representative of the Asian population, and incorporating technical improvements to ensure greater anatomical completeness and accuracy.
Methods
The CVH project implemented several key technical improvements:
- Vascular Perfusion with Colored Gelatin: To enhance the identification of blood vessels, the cadavers were perfused with colored gelatin.
- Whole-Body Cryosectioning: Cadavers were embedded in gelatin and cryosectioned whole. This method, combined with a specialized low-temperature chamber (-25°C), prevented section loss that can occur when bodies are cut into smaller pieces.
- Cryo-Milling at -25°C: Milling was performed at extremely low temperatures (-25°C) to prevent small, delicate structures like teeth, nasal conchae, and articular cartilage from detaching and being lost during the process.
- Improved Milling Machine: A larger milling machine table was used to accommodate a whole embedded body, further reducing the risk of data loss from fragmentation.
Datasets
The paper introduces the Chinese Visible Human (CVH) male and female datasets.
- CVH Male:
- Axial anatomical images: 2518 slices, 3072 x 2048 pixels each (~170 µm resolution).
- Total size: 90.65 Gb.
- CT/MRI data: 512 x 512 pixels (~440 µm resolution).
- CVH Female:
- Axial anatomical images: 3640 slices, 3072 x 2048 pixels each (~170 µm resolution for head, ~340 µm for other regions).
- Total size: 131.04 Gb.
- CT/MRI data: 512 x 512 pixels (~440 µm resolution).
These datasets are available at http://www.chinesevisiblehuman.com.
Key Insights and Ideas
- The Importance of Cryo-Preservation for Detail: The success of the CVH project hinges on the meticulous preservation of anatomical detail. The low-temperature milling and whole-body embedding were crucial for capturing fine structures that are often lost in less rigorous methods.
- Addressing Population Diversity: By creating datasets representative of the Asian population, the CVH project expands the scope of digital human resources beyond existing Caucasian models, enabling more diverse research and applications.
- Technological Synergy for Data Integrity: The combination of advanced embalming, precise cryosectioning, and specialized milling equipment demonstrates how integrated technological approaches are essential for creating high-fidelity anatomical datasets.
- Enabling Advanced Visualization: The datasets are designed for 3D reconstruction and visualization, paving the way for more sophisticated virtual dissection, surgical planning, and anatomical education tools.
Key Results
- The CVH male dataset comprises 2518 anatomical cross-sections with a resolution of approximately 170 µm per slice, totaling 90.65 Gb.
- The CVH female dataset contains 3640 anatomical cross-sections, also at approximately 170 µm resolution for the head, totaling 131.04 Gb.
- The CVH datasets offer higher resolution and greater anatomical completeness compared to the original US VHP, particularly in preserving delicate structures.
- The project successfully created digital models of a typical Chinese male and female, representative of the Asian population.
Limitations and Future Work
- Dataset Size: The datasets are very large (90.65 Gb and 131.04 Gb), which can pose challenges for storage and processing.
- Accessibility: While available via FTP or DVDs, access to the full datasets requires a license agreement from the Third Military Medical University.
- Future Work: The paper mentions the availability of MPEG videos of real-time volume rendering, suggesting ongoing efforts in visualization and accessibility. The development of programmable graphics accelerators for real-time 3D visualization of these large datasets is also highlighted as a significant advancement.
Summary updated