Screen-Tone Look Generation with Luminance-Based Patterning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computer graphics (CG) animated movies struggle to achieve a 'comic-book' look due to issues in pattern projection, gradient handling, printing techniques, and quantization, particularly when objects move or change perspective, leading to distracting patterns and loss of realism.
Innovation Solution
A screen-tone look generator system comprising a hatcher and a thresher that quantizes smooth shading, integrates halftone dots and hatched lines, and adjusts patterns based on object position and luminance to simulate printed artifacts, ensuring flexibility and interaction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If smooth gradations are rendered for shading in CG animated movies, then visual realism is improved, but achieving a comic-book look becomes difficult
Solution Approach 1:
The system dynamically switches between smooth rendering and screen-tone patterns based on the desired artistic style. The hatcher and thresher components are conditionally applied to transform smooth shading into comic-book aesthetics while preserving the underlying 3D geometry and lighting information, enabling versatile style adaptation without sacrificing realism when needed.
Solution Approach 2:
The invention changes the rendering parameters by introducing screen-tone patterns (halftone dots, hatched lines) as an additional layer on top of smooth shading. By controlling the density, size, and distribution of these patterns through luminance-based thresholds, the system transforms the visual appearance from photorealistic to comic-book style while maintaining the original shading information.
2Manufacturing precision
If halftone dots and hatched lines are integrated to simulate printed artifacts, then comic-book authenticity is improved, but pattern distortion during camera movement occurs
Solution Approach 1:
The hatcher component dynamically adjusts the screen-tone patterns based on camera position, object orientation, and surface geometry. The system recalculates pattern placement and distortion in real-time to match the projected view, ensuring that halftone dots and hatched lines remain properly aligned with surface contours even during camera movements and perspective changes.
Solution Approach 2:
The system uses feedback from the 3D scene geometry, camera transform, and lighting calculations to continuously adjust pattern parameters. The hatcher monitors surface normals, curvature, and view angle to modify pattern density and orientation, while the thresher adjusts luminance thresholds to maintain consistent visual appearance across different viewing conditions.
3Adaptability or versatility
If quantization is applied to smooth shading, then comic-book style is achieved, but visual fidelity and realism are reduced
Solution Approach 1:
The invention segments the shading information by separating the base smooth rendering from the overlaid screen-tone patterns. The hatcher creates discrete pattern elements (dots, lines) that are superimposed on the continuous shading, while the thresher segments the luminance range into distinct bands. This segmentation approach enables comic-book style quantization while preserving the underlying high-fidelity shading information for areas where realism is needed.
Solution Approach 2:
The system applies different levels of quantization and pattern density to different regions of the image based on local luminance, surface curvature, and importance. The thresher applies luminance-based thresholding to control where patterns appear and their density, allowing high-fidelity smooth shading to remain visible in critical areas while applying comic-book patterns in less important regions, thus maintaining overall visual fidelity while achieving the desired style.
Data Source
Figure 1
Figure 2A
Figure 2B~2D
AI summary
Generating a screen-tone look image for a video, including: receiving a smooth gradation rendered image and position information of pixels in the rendered image; generating a pattern for each pixel of the pixels in the rendered image using the position information; dividing the rendered image into a fixed number of sections defined by luminance to produce a luminance gradient image; and applying the patterns to the fixed number of sections of the luminance gradient image.