Project Chameleon
Adaptive Camouflage Generation
Update II
What was once a hobby has crossed the line into the business world. In July 2016, The Minifig Co. LLC licensed an emulation version of "Digital Flora" to produce high-quality prints on toy parts. The first product was released in late 2016, and my pattern was used again in May 2018 for a re-release. Both products sold out.
​
​
​
​
​
​
​
​
​
Since the last update, more progress has been made than was ever on this original site. Immense advancements to geometry generation and development practices have come with experience, and from four additional courses involving coding. Overzealous research started my minor in Imaging Science, and I've designed new pattern-creating code based on principles of human vision.
​
My most successful pattern is based on fractally-placed quadrangles. Here are links to a few of my photo tests.
​
​
​
​
​
​
​
​
Since the last update, however, I have developed many other new pattern-creating programs and learned about color measurement and perception.
​
Those four output samples of my randomized programs represent eight months of continual ideas, research, development, frustration and elation. I wrote all of the code that creates these patterns, except for Bresenham's genius line algorithm, a point-inside-polygon boolean method, and the Python library utilized to convert my data into images.
​
Not every pattern was designed with deception as its primary goal: I chose to realize several ideas simply because of the problem-solving challenges along the way. It's extraordinarily exciting to create solutions to problems that nobody else knows the answer to or has attempted to solve before. My proudest coding is the reverse-engineering of real-world computer-generated camouflage patterns, and the image above represents one of my latest breakthroughs.
​
As time advanced, I found myself requiring the functionality of a more advanced system than file-by-file impulse programming. For my own benefit, I've implemented (whether realizing it at the time or not) multi-module integration, a model-view-controller approach, and refactoring of functional but slow methods to decrease computation time.
​
It is my hope to expand my professional stance in the camouflage field, but until then, I seek to learn the best software development and imaging science from not only professors, but employers. To do my best in a dream job as a camoufleur, I must be familiar with sciences and development practices very disparate from the obvious fields, and I look to my future career in computer science as the best possible learning ground.
Try to read every word of this text. It goes to show the purpose of distortion patterns like those I generate: delaying detection, not the near-impossible goal of complete concealment.
​
