Slimmable Compressive Autoencoders for Practical Neural Image Compression
This addresses practical image compression needs by providing variable rate and dynamic adjustment of memory, computational cost, and latency, though it is incremental as it builds on existing neural compression methods.
The paper tackled the problem of neural image compression models being heavy, computationally demanding, and optimized for a single rate by proposing slimmable compressive autoencoders (SlimCAEs) that jointly optimize rate and distortion for different capacities, enabling flexible execution with excellent rate-distortion performance.
Neural image compression leverages deep neural networks to outperform traditional image codecs in rate-distortion performance. However, the resulting models are also heavy, computationally demanding and generally optimized for a single rate, limiting their practical use. Focusing on practical image compression, we propose slimmable compressive autoencoders (SlimCAEs), where rate (R) and distortion (D) are jointly optimized for different capacities. Once trained, encoders and decoders can be executed at different capacities, leading to different rates and complexities. We show that a successful implementation of SlimCAEs requires suitable capacity-specific RD tradeoffs. Our experiments show that SlimCAEs are highly flexible models that provide excellent rate-distortion performance, variable rate, and dynamic adjustment of memory, computational cost and latency, thus addressing the main requirements of practical image compression.