Rewriting geometric rules of a GAN
نویسندگان
چکیده
Deep generative models make visual content creation more accessible to novice users by automating the synthesis of diverse, realistic based on a collected dataset. However, current machine learning approaches miss key element creative process - ability synthesize things that go far beyond data distribution and everyday experience. To begin address this issue, we enable user "warp" given model editing just handful original outputs with desired geometric changes. Our method applies low-rank update single layer reconstruct edited examples. Furthermore, combat overfitting, propose latent space augmentation style-mixing. allows create synthesizes endless objects defined changes, enabling new without burden curating large-scale We also demonstrate can be composed achieve aggregated effects, present an interactive interface through composition. Empirical measurements multiple test cases suggest advantage our against recent GAN fine-tuning methods. Finally, showcase several applications using models, including interpolation image editing.
منابع مشابه
Geometric GAN
Generative Adversarial Nets (GANs) represent an important milestone for effective generative models, which has inspired numerous variants seemingly different from each other. One of the main contributions of this paper is to reveal a unified geometric structure in GAN and its variants. Specifically, we show that the adversarial generative model training can be decomposed into three geometric st...
متن کاملinvestigating the feasibility of a proposed model for geometric design of deployable arch structures
deployable scissor type structures are composed of the so-called scissor-like elements (sles), which are connected to each other at an intermediate point through a pivotal connection and allow them to be folded into a compact bundle for storage or transport. several sles are connected to each other in order to form units with regular polygonal plan views. the sides and radii of the polygons are...
Learning Plan Rewriting Rules
Considerable work has been done to automatically learn domain-specific knowledge to improve the performance of domain independent problem solving systems. However, most of this work has focussed on learning search control knowledge-knowledge that can be used by a problem solving system during search to improve its performance. An alternative approach to improving the performance of domain indep...
متن کاملLearning Plan Rewriting Rules
Planning by Rewriting (PbR) is a new paradigm for efcient high-quality planning that exploits plan rewriting rules and e cient local search techniques to transform an easy-to-generate, but possibly suboptimal, initial plan into a high-quality plan. Despite the advantages of PbR in terms of scalability, plan quality, and anytime behavior, PbR requires the user to de ne a set of domain-speci c pl...
متن کاملTiles, rewriting rules and CCS
In 12] we introduced the tile model, a framework encompassing a wide class of computational systems, whose behaviour can be described by certain rewriting rules. We gathered our inspiration both from the world of term rewriting and of con-currency theory, and our formalism recollects many properties of these sources. For example, it provides a compositional way to describe both the states and t...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: ACM Transactions on Graphics
سال: 2022
ISSN: ['0730-0301', '1557-7368']
DOI: https://doi.org/10.1145/3528223.3530065