Search papers, labs, and topics across Lattice.
This paper introduces LUMI, a tokenizer-agnostic framework for lossless RGB image compression that leverages frozen large language model (LLM) backbones. By replacing traditional pixel-as-text tokenization with a pixel embedding module and incorporating intra-patch position encoding, LUMI maintains spatial structure and achieves competitive compression rates across various image domains. The results demonstrate that LUMI not only unifies the compression process across different tokenizer families but also enhances robustness compared to existing tokenizer-based methods.
LUMI revolutionizes lossless image compression by decoupling it from tokenizer behavior, achieving superior performance with a unified approach across different LLM architectures.
Large language model (LLM)-based lossless image compression methods typically represent pixel data through the native text interface of a pretrained model, converting pixel values into token sequences that the LLM processes through its vocabulary head. This design shows that pretrained language models can provide probability estimates for image coding, but it also couples compression to tokenizer behavior, vocabulary-specific numeric tokens, and model-family-specific adaptation. In this paper, we present LUMI (LLM-based Unified Model-agnostic lossless Image compression), a tokenizer-agnostic framework for lossless RGB image compression with frozen LLM backbones. LUMI replaces pixel-as-text tokenization with a pixel embedding module that maps raw intensity and channel information into the continuous embedding space of the LLM. It further introduces intra-patch position encoding to retain two-dimensional spatial structure after flattening, and uses a 256-way prediction head to produce probabilities over the native pixel alphabet. Only the pixel embedding, position encoding, soft-prefix parameters, and prediction head are trained, while the LLM backbone remains fixed. Experiments on natural, medical, and remote-sensing image benchmarks with LLaMA, Qwen, and Gemma backbones show that LUMI provides a unified interface across tokenizer families, achieves competitive compression rates, and improves cross-domain robustness over tokenizer-based LLM compression baselines. These results formulate LLM-based lossless image compression as pixel-space adaptation of frozen foundation models rather than tokenizer-specific language-symbol modeling.
