---
title: 'Implicit Compression: Function-Based Encoding'
url: https://www.emergentmind.com/topics/implicit-compression
type: topic
---

# Implicit Compression: Function-Based Encoding

Implicit compression is a method of data compression in which signals (such as images, video, volumes, or point clouds) are represented not by storing sampled values or transform coefficients, but by overfitting a compact, continuous neural function—an implicit neural representation (INR)—to the data. The compressed bitstream consists of the learned parameters of this network, potentially after quantization, entropy coding, or further model compression. Decoding then reconstructs the data by evaluating the network on desired coordinates. This paradigm has emerged as a highly general, function-based alternative to classical and latent-space codecs for a wide range of modalities, offering unique advantages in representation flexibility, spatial adaptivity, and direct rate–distortion control [2211.06689][2112.04267][2301.09479].

## 1. Mathematical Formulation of Implicit Compression

A canonical INR is a parameterized function $f_\theta:\mathbb{R}^M\to\mathbb{R}^C$ mapping continuous coordinates $x$ to signal values $y$. In implicit compression, the signal $d(x)$ (e.g., image intensities, voxel occupancy, RGB colors) is approximated as $f_\theta(x)\approx d(x)$ over the domain of interest (pixels, voxels, spatial/temporal/spectral coordinates) [2211.06689][2103.03123][2506.01234][2405.19012]. For multi-dimensional data (images, videos, point clouds), $x$ can encompass spatial, angular, temporal, band, or view indices.

The compression process is thus reduced to learning “per-sample” the network parameters $\theta^*$ that minimize a distortion metric $D(\theta)=\mathbb{E}_x\|f_\theta(x)-d(x)\|^2$. Once optimized, these weights are quantized ($q(\theta)$), perhaps sparsified or pruned, and entropy-coded (using learned, empirical, or parametric priors). The compressed bitstream consists mainly of the coded $\theta$, plus minor metadata [2112.04267][2303.03028][2506.01234][2407.10325].

The fundamental rate–distortion objective is
$$
L(\theta)=D(\theta)+\lambda R(\theta)
$$
where $R(\theta)$ denotes the total coded bit-length and $\lambda$ trades off reconstruction distortion against bitrate [2112.04267][2303.03028][2506.01234].

## 2. Theoretical underpinnings and Spectral Properties

Implicit neural compression leverages specific inductive biases and spectral characteristics of the neural architectures. Shallow MLPs with sinusoidal activation (e.g., SIREN) embedded with Fourier features tend to concentrate approximation power around frequencies represented by their input encoding and hidden-layer width, leading to a “spectrum concentration” effect. This property strongly governs parameter efficiency and reconstruction fidelity for data with broadband or multiscale content [2209.15180].

For a given function $d(x)$ with a target spectrum, the number of INR parameters, the frequency basis of the encoding, and the depth of the network determine how well high-frequency details are preserved. A key finding is that most spectral energy in shallow sinusoidal networks is localized around base frequencies, with amplitude in higher-order harmonics suppressed by the Bessel function decay in the Jacobi–Anger expansion. Thus, signals with broad or spatially-varying spectrum benefit from block-wise partitioning and local models [2209.15180][2211.06689].

## 3. Model Architectures and Compression Workflows

The INR used for implicit compression is usually a compact MLP with periodic (sinusoidal, cosine), ReLU, or learnable activations, often applied after a Fourier feature positional encoding for high-frequency fidelity [2103.03123][2209.15180][2504.14471]. Notable architectural augmentations include

- **Low-rank and adaptive basis factorization** for reducing model size while retaining expressivity (e.g., ImpliSat framework uses low-rank modulated SIREN MLPs for multispectral image compression) [2506.01234].
- **Hierarchical parameter sharing**, as in Tree-structured Implicit Neural Compression (TINC), which encodes both local and non-local redundancy by sharing early layers and only specializing late layers or additive blocks for high-frequency or region-specific details [2211.06689].
- **Block-wise partitioning**, spectrum-based splitting (e.g., SCI), or patch-based optimization, allocating network capacity according to local frequency content or spatial statistics [2209.15180].
- **Learnable and adaptive activation functions** in latent space, such as LeAFNet for point clouds, allowing fine spectral adaptation while maintaining a compact parameterization [2504.14471].

The generic workflow is:

1. Fit an INR $f_\theta$ to the target data sample;
2. Quantize and entropy-code the weights $\theta$ (with bit-precise, possibly layer-wise quantization, entropy coding, or variational models);
3. Transmit the resulting bitstream; and
4. Decode by reconstructing the signal as $f_{\hat\theta}(x)$ on the target grid.

In video and dynamic point-cloud compression, spatio-temporal or view-time coordinates drive the INR, with temporal redundancy handled either via sequential parameterization, residuals, or global 4D embedding [2503.19604][2412.10433]. For multi-spectral images, dynamic Fourier modulation and hypernetworks adapt the parameterization per band and resolution [2506.01234].

## 4. Quantization, Entropy Coding, and Rate–Distortion Control

Compressed INRs are made bit-efficient via aggressive quantization (e.g., down to 7–8 bit precision with minimal degradation using quantization-aware training or AdaRound) [2112.04267][2303.03028], border-aware and empirical entropy coding of quantized weights [2303.03028][2112.04267], and learned priors on parameters (Gaussian, variational, data-driven) [2305.19185][2301.09479].

Some frameworks perform meta-learning (e.g., with MAML) to find a global initialization for quick per-sample fitting and to minimize the amplitude and entropy of weight updates, which facilitates better quantization and entropy coding [2112.04267]. Bayesian approaches fit variational posteriors to parameters and perform entropy coding using relative-entropy codes, setting rate precisely as the KL divergence to the prior ($\mathrm{KL}[q(w)||p(w)]$) [2305.19185].

The overall rate–distortion optimization is sometimes convexified and controlled directly during fitting by Lagrange multipliers or alternating quantization-aware and distortion-reducing gradient steps [2103.03123][2303.03028][2506.01234].

## 5. Applications and Empirical Results Across Modalities

### Images, Light Fields, and Volumetric Data

- Simple INR codecs (COIN, RQAT-INR) can outpace JPEG and approach learned autoencoders at low bitrates, with competitive PSNR and lower decoder complexity [2103.03123][2303.03028].
- Meta-learned and quantization-aware INRs can close the gap to JPEG2000 and BPG at mid rates; INRs trained with entropy coding and advanced architecture outperform block coders on certain benchmarks [2112.04267][2303.03028].
- Light-field INR compression encodes SAIs as functions of (u,v), storing angular redundancy implicitly, and achieves comparable or superior perceptual quality versus classical approaches [2407.10325].

### Biomedical, Climate, and Scientific Data

- Spectrum-concentrated and tree-based INRs (SCI, TINC) achieve state-of-the-art performance on medical volumes, neuronal structures, and atmospheric fields, outperforming HEVC and tailored codecs at high compression ratios (up to 512–1024×) [2211.06689][2209.15180][2411.06155].
- Application-specific guidance (segmentation, perceptual, or task loss) can be seamlessly incorporated during fitting to prioritize critical features in microscopy or scientific imaging [2405.19012].

### Point Cloud and Dynamic Scene Compression

- Per-instance INR fitting for geometry and color fields, with quantization and adaptive activations, surpasses octree and G-PCC standards for both static and dynamic point clouds, with up to +7 dB PSNR improvement and up to 89% bit-rate reduction in dynamic settings [2504.14471][2412.10433].

### Video and Multi-View

- Implicit Neural Video Compression (IPF, GIViC) leverages coordinate-mapped frame prediction or full-GOP (Group of Pictures) functional fitting, outperforming MPEG-4 at low rates and, with generative hierarchical models, surpassing VTM and state-of-the-art neural codecs on benchmark sets [2503.19604][2112.11312].
- Multi-view integration approaches blend implicit and explicit coding, using INR-based codecs to encode all but one anchor view, enabling state-of-the-art scene modeling and improved rate-distortion over MIV [2311.17350].

### Advanced: Quantum and Modality-Agnostic Compression

- Quantum neural architectures (quINR) improve rate-distortion by leveraging exponential Hilbert space capacity, delivering up to +1.2 dB PSNR gain over classical MLP INR on images at a given bit rate [2412.19828].
- Modality-agnostic variational INR compression (VC-INR) yields state-of-the-art performance across images, audio, climate, and 3D scenes, with architecture-agnostic soft-subnetwork gating and learned entropy models for the latent space [2301.09479].

## 6. Limitations, Generalization, and Extensions

The principal limitations of implicit compression are encoder-side computational cost (per-sample fitting generally requires thousands of gradient steps), challenges in capturing high-frequency content with small models (spectral bandwidth bottleneck), and the lack of fast, amortized inference for deployment at scale [2103.03123][2112.04267][2303.03028][2411.06155]. Overfitting-based methods may blur high-frequency, nonstationary structures unless model size or block count is increased, or spectrum-based partitioning is employed [2209.15180].

Recent approaches address these by:

- Incorporating tree- or spectrum-structured block division with hierarchical parameter sharing [2211.06689][2209.15180].
- Utilizing learned optimizers and meta-initializations to reduce fitting time [2405.19012][2112.04267].
- Hybridizing explicit codecs for the dominant portions of data and using INR-based residual coding or adapters for the error, accelerating fitting without sacrificing quality [2405.19012][2311.17350].
- Employing probabilistic and variational compression methods (e.g., variational Bayesian INRs, relative-entropy coding) that provide finer rate-distortion control, improved entropy modeling, and modality-agnostic applicability [2305.19185][2301.09479].

Extensions under active study include neural architecture search for function parameterization, adaptively-learned quantization or pruning schedules, block-wise fitting for massive scenes, hybrid quantum-classical architectures, and expansion to new modalities including hyperspectral, scientific, environmental, and multi-modal data [2412.19828][2506.01234][2411.06155].

## 7. Summary Table of Recent Implicit Compression Approaches

| Model/Framework                    | Domain         | Key Architectural Features                   | Rate-Distortion/Empirical Outcomes                | Reference       |
|------------------------------------|---------------|----------------------------------------------|---------------------------------------------------|-----------------|
| COIN, RQAT-INR                     | Images        | MLP+Fourier features, QAT, AdaRound          | Outperforms JPEG/JPEG2000 at low bpp              | [2103.03123][2303.03028] |
| TINC                               | Volumes       | Tree-based, shared MLP layers                | +1–2 dB PSNR vs. SOTA at 512–1024× compression    | [2211.06689]    |
| SCI                                | Biomedical    | Spectrum-adaptive partition, funnel MLP       | State-of-the-art on CT/sparse data, PSNR, Acc     | [2209.15180]    |
| GIViC, IPF                         | Video         | INR+diffusion/transformer, coordinate flow   | Outperforms VVC/VTM, MPEG-4 at low/RA bitrate     | [2503.19604][2112.11312] |
| ImpliSat                           | Multispectral | Fourier modulation, per-band adaptation       | 10× compression at PSNR=36 dB (Sentinel-2)        | [2506.01234]    |
| PICO, NeRC³                        | Point Clouds  | Learnable activation, joint geometry/color    | +4–7 dB vs. G-PCC, up to 89% BD-BR reduction      | [2504.14471][2412.10433]|
| INIF, HiHa                         | Microscopy, Atmos| Multi-loss INR, harmonic decomposition    | Up to 512× compression, 20–40× compression for 1e-3 RMSE | [2405.19012][2411.06155] |
| VC-INR, COMBINER                   | Multi-modal   | Variational quantization, gating, Bayes prior | Outperforms JPEG2000, MP3, HEVC cross-modality    | [2301.09479][2305.19185]|

Implicit compression thus constitutes a unifying and highly extensible framework for modality-agnostic, function-based compression, capable of rivaling or surpassing expert-tuned codecs in diverse domains through advances in rate-distortion optimization, network architecture, and entropy modeling.

Source: https://www.emergentmind.com/topics/implicit-compression