Idempotence and Perceptual Image Compression

2024-01-17

International Conference on Learning Representations

AI Summary

This work addresses the challenge of directly deploying unconditional generative models for perceptual image compression. We propose a novel idempotence-constrained paradigm: first establishing the theoretical equivalence between idempotence and conditional generative coding, thereby enabling inversion of pre-trained unconditional GANs into perceptual compressors without fine-tuning or retraining. Our method integrates idempotence optimization, GAN inversion, MSE-based end-to-end codec co-optimization, and FID-driven quality assessment. It achieves high-fidelity reconstruction while significantly enhancing perceptual quality—outperforming state-of-the-art methods including HiFiC and ILLM in FID. The framework is theoretically grounded and empirically validated; source code is publicly released to ensure reproducibility and facilitate further research.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

Idempotence is the stability of image codec to re-compression. At the first glance, it is unrelated to perceptual image compression. However, we find that theoretically: 1) Conditional generative model-based perceptual codec satisfies idempotence; 2) Unconditional generative model with idempotence constraint is equivalent to conditional generative codec. Based on this newfound equivalence, we propose a new paradigm of perceptual image codec by inverting unconditional generative model with idempotence constraints. Our codec is theoretically equivalent to conditional generative codec, and it does not require training new models. Instead, it only requires a pre-trained mean-square-error codec and unconditional generative model. Empirically, we show that our proposed approach outperforms state-of-the-art methods such as HiFiC and ILLM, in terms of Fr'echet Inception Distance (FID). The source code is provided in https://github.com/tongdaxu/Idempotence-and-Perceptual-Image-Compression.

Problem

• Idempotence
• Perceptual Image Compression
• Unconditional Generative Models

Innovation

• Perceptual Image Compression
• Idempotent Conditional Generation Models
• FID Score Improvement