# Video Super-Resolution

Project Starlight is an AI video enhancement model launched by Topaz Labs, designed to improve the quality of low-resolution and damaged videos. It uses diffusion model technology to achieve video super-resolution, noise reduction, deblurring, and sharpening, while maintaining temporal consistency to ensure smooth transitions between video frames. This technology is a significant breakthrough in the field of video enhancement, bringing unprecedented high-quality results to video repair and enhancement. Currently, Project Starlight offers a free trial and plans to support 4K export in the future, primarily targeting users and enterprises who need high-quality video repair and enhancement.

STAR is an innovative video super-resolution technology that addresses the issue of over-smoothing found in traditional GAN methods by combining text-to-video diffusion models with video super-resolution. This technology not only recovers video details but also maintains temporal and spatial consistency, making it suitable for various real-world video scenarios. STAR was jointly developed by Nanjing University and ByteDance, boasting high academic value and application prospects.

EvTexture is an event-based, vision-driven video super-resolution (VSR) technology that leverages high-frequency details in event signals to better recover texture regions in VSR. This technique uniquely proposes using event signals for texture enhancement. It iteratively refines texture regions by exploring high temporal resolution event information through a texture enhancement module, achieving more accurate and richer high-resolution details. On four datasets, EvTexture has achieved state-of-the-art performance, particularly on the Vid4 dataset, where it can achieve a gain of up to 4.67dB compared to recent event-based methods.

FMA-Net is a deep learning model specialized in video super-resolution and deblurring. It is designed to restore videos of low resolution and blur into high resolution and clarity. The model achieves this through a combination of flow-guided dynamic filtering and iterative feature refinement using multi-attention techniques, which are effective in handling large motions within the video. This results in a joint super-resolution and deblurring of videos. The model boasts its simplicity in structure and notable effectiveness, making it suitable for wide application in video enhancement and editing fields.