Titan aimed to speed up the innovation process by providing guidance to product designers. Generative adversarial networks (GANs) are a type of generative modeling that can be trained to automatically recognize and learn from patterns in new data. As a promising future option, GANs offer businesses a chance to innovate through fresh product design.
Build a GAN-based image synthesizer in the background that can create new designs from the attributes of existing designs (such as dial color, numbering system (Roman or generic), strap style, etc.).
A GUI that facilitates user engagement with GAN. Give users the opportunity to experiment with a canvas consisting of multiple images. The greater the pointer's proximity to a node image, the greater the resemblance between the two.
Frontend: Mapping coordinates to weightages to correct sequence number from all possible outputs, fetching it from the stream and displaying the output realtime with no timelag
Backend: Ensuring all the combinations are generated with different weightages from different node images and be able to supply the stream real time