Channel and Space-based Joint Rate Allocation Algorithm
Résumé
Rate control is a critical component for image and video compression Particularly under limited network bandwidth conditions, bitrate control is essential to ensure efficient image transmission by effectively allocation channel resources. In this research, since both Channel and Spatial have relationship with rate allocation, we first propose a joint Channel-wise and Spatial-wise Quantization scheme to determine optimal quantization parameters. Subsequently, we develop a quantization step estimation network to obtain parameters to efficiently allocate rate according to target rate. Experiments demonstrate that our algorithm significantly improve compressed image quality with minimal bitrate distortion and achieve accurate rate control with nearly 3% average bitrate error.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|---|
| Licence |