Scalable Coding

Scalable coding aims to create compressed representations of data (e.g., images, video) that can efficiently adapt to varying bandwidth or computational constraints, serving both human perception and machine analysis needs. Current research focuses on improving the rate-distortion performance of these representations, particularly for the machine-analysis component, often leveraging deep neural networks and exploring techniques like conditional and residual coding to better utilize information across different tasks. These advancements are significant for applications such as video surveillance and autonomous systems, where efficient transmission and processing of visual data are crucial.