![]() Thus it is a lengthy and tedious process for both. Each cycle, the designer should set their designs apart from the masses and develop innovative and unique designs for the client to select his favorites from many ultimate drafts produced. During this process, the logo which matches the client refines multiple cycles. Manual design of logos for a new brand is a back-and-forth process between a designer and a client. As a result, their applications to any tailor-made specific problems may still require significant redevelopment. While the existing research on generative image synthesis has achieved significant progress, their development on GANs are primarily limited to the principle of Min–Max game as well as its embedded adversarial learning process. Over the recent years, a range of generative models, ,, ,, ,, ,, ,, ,, ,, , been developed for general image synthesis, where representative state of the arts include texture-based image synthesis, image super-resolution, and photo-realistic facial image synthesis. Extensive experiments are carried out and the results in comparison with the existing state of the arts illustrate that our proposed achieves significant superiority in terms of both synthesized logo quality, integrity and variety. Since learning across different frequency band constantly varies, we further propose a dynamic optimization scheme to maximize the effectiveness of contributions from each individual DCT frequency band. As a result, a new channel of DCT domain generative learning can be established to support the existing pixel domain learning towards improved logo synthesis. To achieve exploitation of all the pixel correlations inside the whole image regardless of their spatial locations, we introduce an approximated DCT transformation and decompose both the input images and the generated images into relatively independent DCT frequency bands. In this paper, we analyze the spectral bias from a frequency perspective to overcome such limitations and hence propose a dynamic self-adaptive optimization on GAN-based generative learning, leading to a dynamic and generative logo synthesis in DCT domain. ![]() Generative learning in pixel domain has achieved great success in exploiting their correlations in processing images towards desired objectives, yet learning in frequency domain could provide added benefits in exploiting pixel correlations without worrying about their spatial locations and increasing their modeling costs.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |