PulseAugur
EN
LIVE 05:13:21

New generative video codecs achieve ultra-low bitrates with high quality · 2 sources tracked

By PulseAugur Editorial · [2 sources] ·

Two research papers introduce novel generative video coding techniques that significantly reduce bitrate while maintaining high visual quality. ReGenVC achieves ultra-low bitrates by compressing video into a compact bitstream of frame data, pose keypoints, and metadata, and reconstructs it using a distilled diffusion transformer. GenVC, another approach, trains a video diffusion model from scratch specifically for compression, employing a global-to-local hierarchy and Adaptive Score Distillation to ensure coherent motion and high perceptual quality. AI

IMPACT These generative video codecs could enable new applications requiring high-quality video transmission at significantly reduced bandwidth.

RANK_REASON Two academic papers published on arXiv detailing new generative video compression models.

Read on arXiv cs.CV →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

New generative video codecs achieve ultra-low bitrates with high quality · 2 sources tracked

COVERAGE [2]

  1. arXiv cs.CV TIER_1 English(EN) · Zheyuan Zhang, Johnson Wu ·

    ReGenVC: End-to-End Real-Time Generative Video Coding at Ultra-Low Bitrate

    arXiv:2607.28144v1 Announce Type: cross Abstract: We present ReGenVC, an end-to-end generative video codec that compresses talking-head video to an ultra-low bitrate and decodes it in real time. The encoder reduces a source clip to a compact bitstream -- a neurally compressed fir…

  2. arXiv cs.CV TIER_1 English(EN) · Naifu Xue, Zhaoyang Jia, Haosen Li, Zihan Zheng, Jiahao Li, Bin Li, Xiaoyi Zhang, Qi Meng, Yuan Zhang, Yan Lu ·

    Generative Video Compression with Adaptive Score Distillation

    arXiv:2607.22772v1 Announce Type: cross Abstract: Diffusion models provide strong generative capabilities for video compression at ultra-low bitrates. Existing diffusion-based video codecs adapt base models originally developed for text-conditioned generation, whereas diffusion m…

RELATED ENTITIES

RELATED TOPICS