A process that converts images into numerical representations that a model can understand and process.
Quality of vision, audio, and image understanding (distinct from modality support)