Bas-Relief Modeling from Normal Images with Intuitive Styles.

Area of Science:

• Computer Graphics
• Digital Art
• Computational Geometry

Background:

• Traditional 3D bas-relief modeling is constrained by model dependency and limited stylistic variations.
• Existing methods often struggle with detail preservation and intuitive control over relief styles.

Purpose of the Study:

• To present a novel method for digital bas-relief modeling with intuitive style control.
• To generate discontinuity-free depth fields with high compression from composite normal images.
• To enable detail enhancement and flexible style design in bas-relief creation.

Main Methods:

• Composing multiple layers of normal images into a single composite normal image.
• Controlling bas-relief style using a parameter and targeted height.
• Solving a sparse linear system for simultaneous bas-relief modeling and stylization.
• Designing bas-reliefs in normal image space using standard image editing tools.

Main Results:

• The proposed method effectively generates digital bas-reliefs with preserved or enhanced fine details.
• It allows for discontinuity-free depth field generation with significant data compression.
• Experiments demonstrate successful application across diverse 3D models and scenes.

Conclusions:

• The novel method overcomes limitations of traditional approaches, offering greater flexibility and control.
• Designing in normal image space democratizes bas-relief creation, making it accessible with common tools.
• This technique advances digital art and 3D modeling by enabling intuitive and detailed relief generation.