Automatic Line Drawing Coloring via Two-Step Model Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional software struggles to automatically color line drawing images effectively, as these images lack luminance information and often have open regions, making it difficult for existing colorization technologies to accurately process and color them.
Innovation Solution
An automatic line drawing coloring method and apparatus that utilize a two-step coloring process, where line drawing data is first reduced in size and then colored using pre-learned models, allowing for the integration of hint information and enabling user selection of models for varied coloring tendencies, and a graphical user interface for displaying and interacting with the coloring process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional software is used for coloring, then closed regions can be colored, but line drawing images with open regions cannot be easily colored
Solution Approach 1:
The system performs preliminary actions by reducing the line drawing image to a smaller size before coloring, and by pre-learning coloring models from sample data. This preparation enables the coloring algorithm to handle open regions effectively, as the reduced image allows better processing of boundary information and region identification.
Solution Approach 2:
The invention changes the parameter of image size by reducing the line drawing image to a predetermined smaller size before coloring. This parameter change enables the coloring algorithm to process open regions more effectively, as the reduced resolution allows better detection and coloring of regions without requiring perfectly closed boundaries.
2Productivity
If black-and-white photographs are colorized using luminance information, then coloring can be achieved, but line drawing images without luminance information are more difficult to process
Solution Approach 1:
The system introduces an intermediary approach by using pre-learned coloring models that were trained on sample line drawing images. These models serve as mediators that capture the relationship between line drawing structures and appropriate colors, enabling the system to colorize images without requiring luminance information or complex real-time analysis.
Solution Approach 2:
The coloring models are prepared in advance through pre-learning from sample data. This preliminary action stores the knowledge of how to translate line drawing structures into colors, allowing rapid coloring of new images without needing to perform complex analysis during the actual coloring process.
3Manufacturing precision
If large-sized line drawing images are processed directly, then original quality is maintained, but processing time and computational load increase
Solution Approach 1:
The system segments the coloring process into two stages: first reducing the image to a smaller size for preliminary coloring processing, then using the colored reduced image to guide the final coloring of the original image. This segmentation allows the computationally intensive coloring algorithm to work on a smaller image initially, reducing overall processing time while maintaining quality.
Solution Approach 2:
The system performs preliminary reduction of the image to a predetermined smaller size before coloring. This preliminary action reduces the computational load for the coloring algorithm, as it processes a smaller image first. The colored reduced image then serves as a guide for coloring the original large image, maintaining quality while reducing processing time.
Data Source
AI summary
An apparatus and a method for coloring line drawing is disclosed for: acquiring line drawing data; performing reduction processing on the line drawing data to be a predetermined reduced size to obtain reduced line drawing data; coloring the reduced line drawing data based on a first learned model which is learned in advance using sample data; and coloring original line drawing data with the colored reduced data and the original line drawing data as inputs based on a second learned model which is learned in advance.