Digitalized Caligraphy:
Individual Work
2020 Summer
Role
Research, Coding
​
Advisor
Fumiya Lida
Tools
Matlab, Python
The research is accomplished through a coding program of Python to focus on generating pixelated typography with the selected Chinese character pattern and to find the beauty within the randomness of Genetic Algorithm.
​
The first proper digital font created by German inventor Rudolf Hell. His CRT-based Digiset machine required digital fonts to be designed using points of light on a grid in bitmap format. The experiment with Chinese digital fonts especially how the computer is writing the singular component character is worth to discover.
During the summer research with the Cambridge Professor Fumiya Lida, to explore the foundations of bio-inspired robotics. The topic of Genetic Algorithm and Game of Life had been discussed, which inspired me to generate digital Chinese typography with reversed Game of life.
Background
1.Core Element of Design- Digital Typography History:
1.1 Photomechanical Process For Printing Fonts:
.png)
.png)
1.2 CRT-based Digiset Process For Printing Digital Fonts:
Variants of the phototypesetting machines utilized the cathode ray tube (CRT) screens to display already digitized characters. The first proper digital font created by German inventor Rudolf Hell. His CRT-based Digiset machine required digital fonts to be designed using points of light on a grid in bitmap format.
.png)
_gif.gif)
3. First Chinese Digitalized Fonts Discovery:
For each Chinese character, designers had to make decisions in the bitmap to preserve calligraphic elegance aesthetics of the calligraphic that was mainly hand-drawn with a brush. One single dislocation in the stroke will result in a bad composition which is difficult to find proper digitalized composition and aesthetics.
How do computation programmers re-create orthographic features of the handwritten Chinese typography, with the different strokes and process within calligraphy?
Single Component Character "Wang": Find the aesthetics within the computation process of writing the character
_gif.gif)
.png)
Within all the pixelated Chinese character, a type of character named single component in which one is derived from a coherent glyph without any division. Different Chinese characters have its unique process for writing. I wonder if there could be a new way in programming this kind of process through algorithm to observe how computation is writing the character.
I finally choose the training character with my last name "Wang", which means king in Chinese which is also one of the largest family name.
Method
Reversing Game of Life Using Genetic Algorithms to Final the Target Chinese Character
In the Bio-inspired Robotics research program. I had amazed by the the cellular automaton that only followed the simple rules from Conway's Game of Life. However, the Chinese character cannot derived from the Conway's rules since the process is reversed. In order to achieve the target character "Wang", the Genetic Algorithm with its metaheuristic reaches optimization through Fitness Function and genetic representation.
.png)
Generate
Initial Population
Fitness Function
for Selection
Mutation
Crossover
20x20 board represents start configuration.
Random seed=50, with 5 iterations
Fitness function selection pick the best individuals to proceed into the next generation.
Introduce random into population for optimization
Original
Mutated
Mother
Father
Child 1
Child 2
Recombine two genes, simulating mating of two children each with half of the genes
.png)
.png)
Generation Matrix &Training Result
Genetic Algorithm coding for each step
.png)
The overall generation process and result of the target Chinese pattern "Wang" is shown below, from the different generation size to the final fitness score. The final 1.00 fitness is able to achieve which means the optimization of the complete pattern when random immigration is added after the generation size of 300.
Result
​
.png)