Skip to main content
GameDev.net gamedev.net

Style Transfer News

The latest Style Transfer coverage curated for game developers.

GLUT: 3D Gaussian Lookup Table for Continuous Color Transformation

For graphics teams, the interesting bit is a continuous color transform that avoids the usual 3D LUT grid trade-off. GLUT uses learnable 3D Gaussian primitives instead of fixed RGB cells, …

Research arXiv cs.GR · 4 months, 2 weeks ago
VQ-Style: Disentangling Style and Content in Motion with Residual Quantized Representations

Disney Research Studios has introduced a groundbreaking method for disentangling style and content in human motion data, which could significantly enhance style transfer in game development. This innovation is particularly …

Research Disney Research Studios · 6 months ago
Parameterized Brushstroke Style Transfer

A new approach to style transfer in game art is emerging, focusing on brush strokes rather than pixels. This method promises to enhance the visual quality of artistic representations, making …

Research arXiv cs.GR · 6 months, 3 weeks ago
RefRetouch: Personalized Image Retouching without Test-time Fine-tuning

RefRetouch tackles personalized image retouching without test-time fine-tuning, aiming to make user-specific style transfer immediate instead of per-user retraining. The system combines an asymmetric auto-encoder with retrieval-augmented retouching to borrow …

Research arXiv cs.GR · 7 months, 2 weeks ago
Untwisting RoPE: Frequency Control for Shared Attention in DiTs

The new insights into Rotary Positional Embeddings (RoPE) could significantly impact how artists and designers approach style transfer in generative models. By modulating frequency bands, developers can achieve a better …

Research arXiv cs.GR · 7 months, 4 weeks ago
Thinking Like Van Gogh: Structure-Aware Style Transfer via Flow-Guided 3D Gaussian Splatting

A new framework for 3D style transfer, inspired by Van Gogh's artistic principles, emphasizes geometric abstraction over traditional texture projection. This approach could significantly enhance the capabilities of graphics programmers …

Research arXiv cs.GR · 8 months, 2 weeks ago
LouvreSAE: Sparse Autoencoders for Interpretable and Controllable Style Transfer

LouvreSAE introduces a novel approach to artistic style transfer that significantly reduces computational overhead. By utilizing a Sparse Autoencoder, developers can achieve style transfer without the need for model fine-tuning …

Research arXiv cs.GR · 9 months, 1 week ago