High-Fidelity Compression of Seismic Velocity Models via SIREN Auto-Decoders

TL;DR

This paper introduces a high-fidelity seismic velocity model compression method using SIREN auto-decoders, enabling efficient storage, smooth interpolation, and super-resolution without retraining, with promising results on diverse geological data.

Contribution

The paper presents a novel INR-based auto-decoder framework for compressing seismic velocity models, achieving high-quality reconstruction and zero-shot super-resolution capabilities.

Findings

Achieves 19:1 compression ratio with high PSNR and SSIM.

Enables smooth latent space interpolation for plausible intermediate models.

Supports zero-shot super-resolution up to 280x280 resolution.

Abstract

Implicit Neural Representations (INRs) have emerged as a powerful paradigm for representing continuous signals independently of grid resolution. In this paper, we propose a high-fidelity neural compression framework based on a SIREN (Sinusoidal Representation Networks) auto-decoder to represent multi-structural seismic velocity models from the OpenFWI benchmark. Our method compresses each 70x70 velocity map (4,900 points) into a compact 256-dimensional latent vector, achieving a compression ratio of 19:1. We evaluate the framework on 1,000 samples across five diverse geological families: FlatVel, CurveVel, FlatFault, CurveFault, and Style. Experimental results demonstrate an average PSNR of 32.47 dB and SSIM of 0.956, indicating high-quality reconstruction. Furthermore, we showcase two key advantages of our implicit representation: (1) smooth latent space interpolation that generates plausible intermediate velocity structures, and (2) zero-shot super-resolution capability that reconstructs velocity fields at arbitrary resolutions up to 280x280 without additional training. The results highlight the potential of INR-based auto-decoders for efficient storage, multi-scale analysis, and downstream geophysical applications such as full waveform inversion.