Monday, March 8, 2010

Computer Aided Fashion Design

Fashion is a big industry. Billions of dollars are thrown around to get just the right look. And for now, they all have to be designed by hand. Computers and sewing technology may help assemble the final product, but it starts with a hand drawn sketch.
But, an important invention in Computer Science and Biology will, if applied here, assist the fashion designer greatly. I speak, of course, of genetic algorithms.
Evolution is an important principle in modern biology, and the subject is frankly impossible to understand without it. Evolutionary biology states that genes drift over time, and successful ones predominate over unsuccessful. Therefore, life adjusts to current conditions, whatever they may be.
Genetic algorithms adjust computer code, or abstract computerized representations, according to a fitness function. When the genetic algorithm is trying to design something, the fitness function works according to the qualities of the thing trying to be designed. When the genetic algorithm is designing an algorithm, the fitness function determines how efficiently the target runs.
The algorithm starts with a pool of different designs. Each are tested. A fraction of the ones that perform the least well are removed from the pool, and are now "dead." The same proprtion of those that perform the most well are combined with each other to produce new designs to take their place. A different fraction gets "mutated" with small variations on their design. Then the entire process runs again.
The genetic fashion design algorithm will start with a representation of a person (almost assuredly a woman, due to stereotyping in the industry and culture), and 25 clothing designs, copied from previous years fashions.. The fashion designer will be asked to click the one that appeals to him (or her) the most. One design would randomly be removed. Designs would have to retain a number of clicks every round to "survive." After 7 rounds, all designs that had not received at least one click would "die." After 15 rounds, two clicks. The mutation rate shouldn't exceed two mutations per round.
A good genetic algorithm typically requires thousands of rounds, but I don't anticipate this taking more than two hours. And after it's done, one design will prevail that the fashion designer absolutely adores. He (or she) will probably put a personal touch on top when he (or she) draws it, producing something probably better than would be made either by computer or human alone.
Alternatively, this could be made into a Flash application that random samples of the Internet could mess with, the results being statistically tallied and sent to the design room, as a "random sample" of what people think about fashion. Fashion designers probably shouldn't rely on this, as people have been known to crash online polls "for the lulz" or to push some point that the public at large doesn't agree with.

No comments:

Related Posts Plugin for WordPress, Blogger...