H/t: Peta Pixel
Max Planck Institute for Informatics researchers have collaborated wit the University of Hong Kong to develop StyleNeRF, a 3D-aware generative model trained on unstructured 2D images capable of synthesizing photorealistic images with a high level of multi-view consistency. Current approaches either cannot synthesize high-resolution images with fine details or yield clearly noticeable 3D-inconsistent artifacts, this approach tackles these issues by integrating the neural radiance field (NeRF) into a style-based generator.
To address the first issue of synthesizing high-resolution images with fine details, the research team performed volume rendering only to produce a low-resolution feature map, and then progressively applied up-sampling in 2D. To lessen the inconsistencies caused by 2D up-sampling, multiple designs were proposed, including a better up-sampler choice and a new regularization loss to enforce 3D consistency. Using these new designs, StyleNeRF was able to synthesize high-resolution images at interactive rates while preserving 3D consistency at exceptional quality.
- Powerful Productivity: AMD Ryzen 3 3350U delivers desktop-class performance and amazing battery life in a slim notebook. With Precision Boost, get up...
- Maximized Visuals: See even more on the stunning 15.6" Full HD display with 82.58% screen-to-body, 16:9 aspect ratio and narrow bezels
- Backlit Keyboard and Fingerprint Reader: Biometric fingerprint reader and Windows Hello sign-in options help keep your Acer PC secure
StyleNeRF also enables control of camera poses and different levels of styles, which can generalize to unseen views. It also supports challenging tasks such as style mixing, inversion and simple semantic edits,” said the researchers.
