GRAM: Generative Radiance Manifolds for 3D-Aware Image Generation

Yu, Dao‐Yi; Yang, Jiaolong; Xiang, Jianfeng; Tong, Xin

doi:10.48550/arxiv.2112.08867

Cited by 4 publications

(4 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 3 The results of PICLANony and face generation methods(CONFIG (Kowalski et al (2020)), StyleGAN2 (Karras et al (2020)), GRAM (Deng et al (2022)), AniFaceGAN (Wu et al (2022)), Pof3d (Shi et al (2023))) on FID and KID. Red is the best, followed by bold.…”

Section: Quantitative Resultsmentioning

confidence: 99%

PICLAnony: Controlled Anonymization of Face Attributes Based on Parametric Imitation Comparison Learning

Liu,

Ji,

Liang

2024

Preprint

View full text Add to dashboard Cite

With the ubiquitous use of mobile devices and the widespread sharing of personal photos and videos on social media, the use of anonymized images to avoid identity disclosure has become essential. Aiming at the problem of low-quality and uncontrollable anonymous faces generated by existing anonymization algorithms, we propose PICLAnony, a controllable anonymization algorithm for face attributes based on parametric imitation comparison learning. It transfers the four visual information corresponding to identity, expression, pose, and illumination of the source image to the generated anonymized face image through parametric imitation contrast learning. And it edits these attribute features that reflecting sensitive behavioral intentions in a context-controlled manner. In the parameter imitation learning stage, high-quality and pose-controllable anonymized faces are generated by imitating the semantic parameters of real images. In the parameter comparison learning stage, the semantic parameters of the edited anonymized image are compared and learned with those of the source image, which solves the problem of insufficient decoupling of expression and illumination attributes in the editing process. In addition, a background control module is designed to keep the background controllable during the editing process of anonymous face facial attributes. We show the subjective results of our algorithm on both CelebA\_HQ and FFHQ datasets, and the subjective and objective results demonstrate that PICLAnony outperforms the state-of-the-art methods in terms of image quality and editing of facial attributes of anonymized faces.

show abstract

Section: Quantitative Resultsmentioning

confidence: 99%

PICLAnony: Controlled Anonymization of Face Attributes Based on Parametric Imitation Comparison Learning

Liu,

Ji,

Liang

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…However, the costly sampling and MLP queries in neural volume rendering make those unsuitable for training high-resolution GANs. To this end, the recently developed 3D-aware GANs make use of a two-stage rendering process [6,17,27,29,42,43] or efficient sampling strategy [14,44] to tackle this problem. Meanwhile, aimed to reduce view-inconsistent artifacts brought by the 2D renderers, they adopt different strategies including NeRF path regularization [17], dual discriminators [6], etc.…”

Section: Related Work 21 Generative 3d-aware Image Synthesismentioning

confidence: 99%

IDE-3D: Interactive Disentangled Editing for High-Resolution 3D-aware Portrait Synthesis

Sun¹,

Wang²,

Shi³

et al. 2022

Preprint

View full text Add to dashboard Cite

“…Concurrently, StyleNeRF [21], CIPS-3D [71], StyleSDF [42] adopt the two-stage rendering strategy to reduce the computation for high-resolution image generation. EG3D [10] introduces tri-plane representation for fast and scalable rendering, and GRAM [16] proposes to render radiance manifolds first to produce high quality images. However, all these concurrent methods lack controllable relighting capabilities.…”

Section: Related Workmentioning

confidence: 99%