Image Generating Apparatus Contour Line Rendering Overlap
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In three-dimensional computer graphics, especially in video games, the realistic rendering of objects can lead to unnatural images when contour lines are drawn for tuned objects overlapping with non-tuned objects, causing disordered and unnatural visual effects.
Innovation Solution
An image generating apparatus that uses a virtual camera to differentiate between tuned and non-tuned objects by determining overlap and selectively drawing contour lines only for the tuned objects, ensuring accurate representation and preventing contour lines from being drawn on non-tuned objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If contour lines are drawn for all tuned objects using standard contour detection, then the contour detection process is simple and fast, but contour lines are incorrectly drawn on non-tuned objects when they overlap with tuned objects, causing unnatural images
Solution Approach 1:
The patent segments the contour detection process into two distinct stages: first detecting contours for all objects without differentiation, then performing a second pass to identify and remove contour lines that appear on non-tuned objects by analyzing depth information and object type data. This segmentation allows the system to maintain fast initial detection while adding precision through the subsequent refinement step.
Solution Approach 2:
The patent introduces depth information and object type identification as intermediary data structures that mediate between the raw contour detection results and the final rendered image. These intermediaries carry information about which objects are tuned versus non-tuned, allowing the system to selectively apply contour lines only where appropriate without sacrificing detection speed.
2Manufacturing precision
If the system differentiates between tuned and non-tuned objects to prevent incorrect contour lines, then contour line accuracy improves, but the system complexity increases due to additional processing steps
Solution Approach 1:
The patent performs preliminary classification of objects into tuned and non-tuned categories before the contour rendering stage. By pre-computing and storing object type information and depth data during the scene setup phase, the system avoids complex real-time differentiation during rendering, thus improving accuracy without proportionally increasing processing complexity.
Solution Approach 2:
The patent applies different processing qualities to different parts of the image: full contour detection is applied to tuned objects while non-tuned objects receive a simplified processing path that excludes contour line drawing. This local differentiation allows the system to maintain high accuracy where needed while reducing overall computational complexity by avoiding unnecessary processing for non-tuned objects.
3Shape
If contour lines are drawn following detected boundaries, then the original shape of objects is preserved, but unnatural images result when tuned and non-tuned objects overlap
Solution Approach 1:
The patent implements a feedback mechanism where the system first draws contour lines based on detected boundaries, then analyzes the resulting image to identify contour lines that appear on non-tuned objects, and finally removes those specific contour lines. This feedback loop ensures that object shapes are faithfully preserved while eliminating the unnatural artifacts caused by incorrect contour placement.
Solution Approach 2:
The patent converts the harmful effect of overlapping objects causing incorrect contour lines into a benefit by using the overlap detection itself as a criterion for selective contour removal. The system identifies where contours appear on non-tuned objects and uses this information to intelligently remove only those specific contour segments, thereby maintaining shape fidelity for tuned objects while eliminating unnatural appearances in overlapping regions.
Data Source
AI summary
The object is to generate an image in which an overlapped part of a tuned object and a non-tuned object having a contour line drawn therein is naturally represented. Only tuned objects among objects existing in a virtual three-dimensional space are perspective transformed, and image data and depth data of each tuned object is written into a process buffer 163a for each pixel. The contour of each tuned object is detected based on the depth data and normal line data acquired from the perspective transformation process, and contour line data is written into a process buffer 163c together with the depth data. When a non-tuned objects is perspective transformed, in a pixel in which a non-tuned objected is located on the front side of a tuned object, image data and depth data of the process buffer 163a are updated. The contour line data of the process buffer 163c is combined with the image data of the process buffer 163a, and the composed data is written into a frame buffer 112. However, the contour line data of pixels having different depth data is not written.