A Taxonomy of Miscompressions: Preparing Image Forensics for Neural Compression
This addresses the challenge of detecting and managing semantic errors in neural compression for applications like forensics, though it is incremental as it focuses on taxonomy rather than solutions.
The paper tackles the problem of neural image compression compromising semantic fidelity despite high perceptual quality, proposing a taxonomy of miscompressions to categorize errors and facilitate risk communication and mitigation research.
Neural compression has the potential to revolutionize lossy image compression. Based on generative models, recent schemes achieve unprecedented compression rates at high perceptual quality but compromise semantic fidelity. Details of decompressed images may appear optically flawless but semantically different from the originals, making compression errors difficult or impossible to detect. We explore the problem space and propose a provisional taxonomy of miscompressions. It defines three types of 'what happens' and has a binary 'high impact' flag indicating miscompressions that alter symbols. We discuss how the taxonomy can facilitate risk communication and research into mitigations.