Exploiting connectivity to improve the tangential part of geometry prediction

Area of Science:

• Computer Graphics
• Geometric Modeling
• Data Compression

Background:

• Efficient compression of triangular meshes is crucial for applications handling large 3D datasets.
• Existing methods primarily focus on normal prediction, often overlooking tangential data behavior.

Purpose of the Study:

• To investigate the impact of vertex valence information on improving tangential prediction accuracy in mesh compression.
• To propose a novel weighted parallelogram prediction method utilizing vertex valences.

Main Methods:

• Incorporation of vertex valence data into prediction algorithms.
• Development of a weighted parallelogram prediction strategy.
• Integration of the proposed method into existing compression frameworks like Edgebreaker.

Main Results:

• Demonstrated enhanced tangential prediction accuracy by considering vertex valences.
• Showcased the adaptability of the method to varying levels of computational complexity.
• Achieved improved compression efficiency compared to traditional methods.

Conclusions:

• Vertex valence information offers a valuable, yet underutilized, feature for enhancing tangential prediction in mesh compression.
• The proposed weighted parallelogram prediction is a computationally efficient and effective enhancement for existing algorithms.
• This approach offers a flexible solution for improving compression performance across different sophistication levels.