Abstract:Art has long been a medium for individuals to engage with the world. Scribble art, a form of abstract visual expression, features spontaneous, gestural strokes made with pens or brushes. These dynamic and expressive compositions, created quickly and impulsively, reveal intricate patterns and hidden meanings upon closer inspection. While scribble art is often associated with spontaneous expression and experimentation, it can also be planned and intentional. Some artists use scribble techniques as a starting point for their creative process, exploring the possibilities of line, shape, and texture before refining their work into more polished compositions. From ancient cave paintings to modern abstract sketches and doodles, scribble art has evolved with civilizations, reflecting diverse artistic movements and cultural influences. This evolution highlights its universal appeal, transcending language and cultural barriers and connecting people through the shared experience of creating art.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to convert images into meaningful scribble art with an artistic sense. Specifically, the authors explored how to use metaheuristics to generate scribble art in order to make up for the deficiencies of existing deep - generation networks in generating scribble art.
### Problem Background
Existing deep - learning methods such as Generative Adversarial Networks (GANs) mainly focus on generating line drawings, contour maps, and edge maps, etc., but have not been fully explored in generating scribble art. Scribble art is an abstract art form, usually consisting of randomly but strategically placed lines, which can capture the structural features of an image while retaining a certain degree of randomness and artistic sense. Therefore, how to generate such scribble art from a given image is a problem worthy of research.
### Research Objectives
The authors hope to generate scribble art through metaheuristics, which can provide better adaptability and accuracy while maintaining computational efficiency. Unlike learning - based methods, metaheuristics do not require a large amount of supervised training data, but gradually find the optimal solution through the optimization process.
### Specific Problems
1. **How to generate meaningful scribble art from an image**: Scribble art needs to introduce randomness and an artistic sense while maintaining the structural features of the image. This requires the algorithm to be able to find the optimal combination of lines in a complex solution space.
2. **How to balance randomness and orderliness**: Scribble art should have both randomness and maintain a certain structure and recognizability. Therefore, the algorithm needs to find a suitable balance between the two.
3. **How to evaluate the quality of the generated scribble art**: In order to ensure that the generated scribble art is both beautiful and has an artistic sense, the authors proposed using indicators such as Peak Signal - to - Noise Ratio (PSNR) and Mean Squared Error (MSE) to evaluate the quality of the generated results.
### Solutions
The authors proposed a framework based on metaheuristics, including Genetic Algorithm (GA), Differential Evolution (DE), Particle Swarm Optimization (PSO), Gravitational Search Algorithm (GSA) and Harris Hawks Optimization (HHO), etc., and modified and optimized these algorithms to make them suitable for the task of generating scribble art. Through experimental verification, these algorithms perform well in generating scribble art and can generate scribble images with an artistic sense and clear structure.
### Conclusions
This research shows the potential of metaheuristics in generating scribble art, especially in the case of lacking a large amount of supervised training data, these algorithms can provide better adaptability and accuracy. Future research can further optimize the parameters of these algorithms, expand their application scope, and explore more diverse art - style generation.
### Related Formulas
- **Mean Squared Error (MSE)**:
\[
\text{MSE}=\frac{1}{N}\sum_{i = 1}^{N}(I_i - I'_i)^2
\]
where \(I_i\) is the pixel value of the input image, \(I'_i\) is the pixel value of the generated scribble image, and \(N\) is the total number of pixels.
- **Peak Signal - to - Noise Ratio (PSNR)**:
\[
\text{PSNR}=10\log_{10}\left(\frac{\text{MAX}_I^2}{\text{MSE}}\right)
\]
where \(\text{MAX}_I\) is the maximum pixel value of the image.
These formulas are used to evaluate the similarity and quality between the generated scribble art and the original image.