Sibling Neural Estimators: Improving Iterative Image Decoding with Gradient Communication
This work addresses image compression quality for applications like storage and transmission, representing an incremental improvement over prior neural methods.
The paper tackles the problem of lossy image compression by developing a neural-based system that uses two recurrent networks communicating via gradients to improve iterative image decoding, resulting in up to 1.64 dB gain over JPEG and 0.37 dB gain over a single iterative neural decoder on the Kodak dataset.
For lossy image compression, we develop a neural-based system which learns a nonlinear estimator for decoding from quantized representations. The system links two recurrent networks that \help" each other reconstruct same target image patches using complementary portions of spatial context that communicate via gradient signals. This dual agent system builds upon prior work that proposed the iterative refinement algorithm for recurrent neural network (RNN)based decoding which improved image reconstruction compared to standard decoding techniques. Our approach, which works with any encoder, neural or non-neural, This system progressively reduces image patch reconstruction error over a fixed number of steps. Experiment with variants of RNN memory cells, with and without future information, find that our model consistently creates lower distortion images of higher perceptual quality compared to other approaches. Specifically, on the Kodak Lossless True Color Image Suite, we observe as much as a 1:64 decibel (dB) gain over JPEG, a 1:46 dB gain over JPEG 2000, a 1:34 dB gain over the GOOG neural baseline, 0:36 over E2E (a modern competitive neural compression model), and 0:37 over a single iterative neural decoder.