Google has announced an experimental project in conjunction with European e-commerce company Zalando that will enable users to create virtual 3D fashion designs.
After acquiring the needed information from the users with a series of questions about their favorite music, art and current mood, the system will use an algorithm to design clothes that matches the user’s personality. The program learned to connect those preferences back to people with similar interests.
This is an effort to show the machine’s process of learning and demonstrate its potential. Google has supported the launching of a new machine-learning system called TensorFlow that made possible the creation of Project Muze.
The program obtained its data from the preferences about color, texture and style of more than 600 fashion experts. Project Muze is an attempt to build a neural network capable of making creative decisions resulting in virtual designs that will be transformed into real-life clothing.
The development in the field of A.I. has found many applications such as the predictive typing keyboard company SwiftKey and the image-recognition used by Facebook. Now, it has entered the world of fashion.
Achim Rietze, from the Google’s ZOO team, says that this experience “It’s like being the muse for your own clothing — your personality and interests become the inspiration for unique designs”.
However, the creations of Project Muze are extremely odd, if not more so. You can get some elements of the outfit that stay apparently suspended by invisible strings. Also, the patterns on the clothing move and flow, as animations — something real-life clothing can’t mimic.
Of course, some of the items could be considered as interesting runway pieces, but in terms of creating real-world, wearable clothing, machines are not quite there yet.
Project Muze is debuting this week at a Berlin trend show called Bread&Butter. All the attendance will be able to participate in the experiment through tablets and on a big screen.