This paper proposes a unified framework that harnesses the synergies between learned image compression (LIC) and zero-shot multi-modal retrieval to enable efficient storage, retrieval, and cross-modal search of multimedia data.
