• Corpus ID: 232110691

COIN: COmpression with Implicit Neural representations

@article{Dupont2021COINCW,
  title={COIN: COmpression with Implicit Neural representations},
  author={Emilien Dupont and Adam Goli'nski and Milad Alizadeh and Yee Whye Teh and A. Doucet},
  journal={ArXiv},
  year={2021},
  volume={abs/2103.03123}
}
We propose a new simple approach for image compression: instead of storing the RGB values for each pixel of an image, we store the weights of a neural network overfitted to the image. Specifically, to encode an image, we fit it with an MLP which maps pixel locations to RGB values. We then quantize and store the weights of this MLP as a code for the image. To decode the image, we simply evaluate the MLP at every pixel location. We found that this simple approach outperforms JPEG at low bit-rates… 

Figures from this paper

Implicit Neural Video Compression
TLDR
The method, which is called implicit pixel flow (IPF), offers several simplifications over established neural video codecs: it does not require the receiver to have access to a pretrained neural network, does not use expensive interpolation-based warping operations, anddoes not require a separate training dataset.
Implicit Neural Representations for Image Compression
TLDR
This work proposes the first comprehensive image compression pipeline based on INRs including quantization, quantization-aware retraining and entropy coding, and finds that the approach to source compression with INRs vastly outperforms similar prior work, is competitive with common compression algorithms designed specifically for images and closes the gap to state-of-theart learned approaches based on Rate-Distortion Autoencoders.
Neural Weight Step Video Compression
TLDR
This work suggests a set of experiments for testing the feasibility of compressing video using two architectural paradigms, coordinate-based MLP (CbMLP) and convolutional network, and proposes a novel technique of neural weight stepping, where subsequent frames of a video are encoded as low-entropy parameter updates.
Neural Distance Oracle for Road Graphs Bachelor ’ s Thesis
TLDR
This thesis focused on distance prediction on MLPs with a simple model structure and tried to reason why some models do well on some types of graphs, while they fail on others and why they sometimes fail on larger graph instances of the same type.
From data to functa: Your data point is a function and you can treat it like one
TLDR
It is demonstrated that the proposed framework for deep learning on functa has various compelling properties across data modalities, in particular on the canonical tasks of generative modeling, data imputation, novel view synthesis and classification.
From data to functa: Your data point is a function and you should treat it like one
TLDR
This paper refers to the data as functa, and proposes a framework for deep learning on funCTa, which has various compelling properties across data modalities, in particular on the canonical tasks of generative modeling, data imputation, novel view synthesis and classification.
Self-Similarity Priors: Neural Collages as Differentiable Fractal Representations
TLDR
This work designs a novel class of implicit operators, Neural Collages, which represent data as the parameters of a self–referential, structured transformation, and employ hypernetworks to amortize the cost of amortizing these parameters to a single forward pass.
COIN++: Data Agnostic Neural Compression
TLDR
This paper proposes COIN++, a neural compression framework that seamlessly handles a wide range of data modalities, and quantize and entropy code these modulations, leading to large compression gains while reducing encoding time by two orders of magnitude compared to baselines.
Optical Flow Regularization of Implicit Neural Representations for Video Frame Interpolation
TLDR
It is shown that constraining the INR derivatives not only allows to better interpolate intermediate frames but also improves the ability of narrow networks to fit the observed frames, which suggests potential applications to video compression and INR optimization.
KRNet: Towards Efficient Knowledge Replay
TLDR
A novel and efficient knowledge recording network (KRNet) which directly maps an arbitrary sample identity number to the corresponding datum and can be trained without the encoder sub-network is proposed.
...
...

References

SHOWING 1-10 OF 39 REFERENCES
Generative Models as Distributions of Functions
TLDR
This paper abandon discretized grids and instead parameterize individual data points by continuous functions, which allows the model to learn rich distributions of functions independently of data type and resolution.
Overfitting for Fun and Profit: Instance-Adaptive Data Compression
TLDR
This paper adapts the full model to a single video, and sends model updates (quantized and compressed using a parameter-space prior) along with the latent representation, and demonstrates that full-model adaptation improves RD performance by ∼ 1 dB, with respect to encoder-only finetuning.
Learned Initializations for Optimizing Coordinate-Based Neural Representations
TLDR
Standard meta-learning algorithms are proposed to be applied to learn the initial weight parameters for fully-connected coordinate-based neural representations based on the underlying class of signals being represented, enabling faster convergence during optimization and resulting in better generalization when only partial observations of a given signal are available.
CompressAI: a PyTorch library and evaluation platform for end-to-end compression research
TLDR
CompressAI is presented, a platform that provides custom operations, layers, models and tools to research, develop and evaluate end-to-end image and video compression codecs and is intended to be soon extended to the video compression domain.
Variable Rate Image Compression with Content Adaptive Optimization
TLDR
A variable rate image compression framework for low bit-rate image compression task that is able to achieve continuously variable rate in a single model by introducing a pair of gain units into VAE.
Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains
TLDR
An approach for selecting problem-specific Fourier features that greatly improves the performance of MLPs for low-dimensional regression tasks relevant to the computer vision and graphics communities is suggested.
MetaSDF: Meta-learning Signed Distance Functions
TLDR
This work formalizes learning of a shape space as a meta-learning problem and leverage gradient-based meta- learning algorithms to solve this task and demonstrates that the proposed gradient- based method outperforms encoder-decoder based methods that leverage pooling-based set encoders.
Implicit Neural Representations with Periodic Activation Functions
TLDR
This work proposes to leverage periodic activation functions for implicit neural representations and demonstrates that these networks, dubbed sinusoidal representation networks or Sirens, are ideally suited for representing complex natural signals and their derivatives.
Equivariant Neural Rendering
TLDR
This work proposes a framework for learning neural scene representations directly from images, without 3D supervision, and introduces a loss which enforces equivariance of the scene representation with respect to 3D transformations.
Improving Inference for Neural Image Compression
TLDR
This work identifies three approximation gaps which limit performance in the conventional approach to compression and proposes improvements to each based on ideas related to iterative inference, stochastic annealing for discrete optimization, and bits-back coding, resulting in the first application of bits- back coding to lossy compression.
...
...