On the Choice of Perception Loss Function for Learned Video Compression
May 30, 2023 Β· Declared Dead Β· π Neural Information Processing Systems
"No code URL or promise found in abstract"
Evidence collected by the PWNC Scanner
Authors
Sadaf Salehkalaibar, Buu Phan, Jun Chen, Wei Yu, Ashish Khisti
arXiv ID
2305.19301
Category
eess.IV: Image & Video Processing
Cross-listed
cs.CV,
cs.IT,
cs.LG
Citations
17
Venue
Neural Information Processing Systems
Last Checked
4 months ago
Abstract
We study causal, low-latency, sequential video compression when the output is subjected to both a mean squared-error (MSE) distortion loss as well as a perception loss to target realism. Motivated by prior approaches, we consider two different perception loss functions (PLFs). The first, PLF-JD, considers the joint distribution (JD) of all the video frames up to the current one, while the second metric, PLF-FMD, considers the framewise marginal distributions (FMD) between the source and reconstruction. Using information theoretic analysis and deep-learning based experiments, we demonstrate that the choice of PLF can have a significant effect on the reconstruction, especially at low-bit rates. In particular, while the reconstruction based on PLF-JD can better preserve the temporal correlation across frames, it also imposes a significant penalty in distortion compared to PLF-FMD and further makes it more difficult to recover from errors made in the earlier output frames. Although the choice of PLF decisively affects reconstruction quality, we also demonstrate that it may not be essential to commit to a particular PLF during encoding and the choice of PLF can be delegated to the decoder. In particular, encoded representations generated by training a system to minimize the MSE (without requiring either PLF) can be {\em near universal} and can generate close to optimal reconstructions for either choice of PLF at the decoder. We validate our results using (one-shot) information-theoretic analysis, detailed study of the rate-distortion-perception tradeoff of the Gauss-Markov source model as well as deep-learning based experiments on moving MNIST and KTH datasets.
Community Contributions
Found the code? Know the venue? Think something is wrong? Let us know!
π Similar Papers
In the same crypt β Image & Video Processing
R.I.P.
π»
Ghosted
π
π
The Cartographer
Deep Learning for Hyperspectral Image Classification: An Overview
R.I.P.
π»
Ghosted
U-Net and its variants for medical image segmentation: theory and applications
R.I.P.
π»
Ghosted
Algorithm Unrolling: Interpretable, Efficient Deep Learning for Signal and Image Processing
R.I.P.
π
404 Not Found
Lightweight Image Super-Resolution with Information Multi-distillation Network
R.I.P.
π»
Ghosted
Deep Learning on Image Denoising: An overview
Died the same way β π» Ghosted
R.I.P.
π»
Ghosted
Federated Learning: Strategies for Improving Communication Efficiency
R.I.P.
π»
Ghosted
In-Datacenter Performance Analysis of a Tensor Processing Unit
R.I.P.
π»
Ghosted
Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning
R.I.P.
π»
Ghosted