HyPER-GAN enhances game engine photorealism with real-time performance

Researchers have developed HyPER-GAN, a novel hybrid image-to-image translation framework designed for real-time photorealism enhancement in game engines. This lightweight U-Net-style generator utilizes a hybrid training strategy, incorporating patches from unpaired real-world images to improve content preservation and visual realism. HyPER-GAN demonstrates a significant performance increase, achieving a 6x speedup at 1080p compared to existing methods, while maintaining temporal consistency and semantic integrity. AI

IMPACT This research could enable more realistic graphics in real-time game engines and simulation applications.

RANK_REASON The cluster describes a new research paper detailing a novel generative model for image-to-image translation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• arXiv
• HyPER-GAN
• Stefanos Pasios

• paper
• product
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →