PerCoV2: Ultra-Low Bit-Rate Perceptual Image Compression via Query-Based 1D Multimodal Image Tokens
PerCoV2 is a novel ultra-low bitrate perceptual image compression system built on Stable Diffusion 3. It enhances efficiency by explicitly modeling the discrete hyper-latent image distribution, allowing it to achieve superior image fidelity at extremely low bitrates while maintaining competitive perceptual quality compared to other methods.