Pruning AMR: Efficient Visualization of Implicit Neural Representations via Weight Matrix Analysis

TL;DR

PruningAMR is a novel algorithm that analyzes and prunes implicit neural representations to generate memory-efficient, adaptive meshes for visualization, without needing access to training data.

Contribution

It introduces a weight matrix analysis method for pruning INRs to facilitate adaptive mesh generation tailored to geometric features.

Findings

Enables automatic, resolution-adaptive mesh generation from pre-trained INRs.

Achieves significant memory savings in visualization tasks.

Operates without access to original training data.

Abstract

An implicit neural representation (INR) is a neural network that approximates a spatiotemporal function. Many memory-intensive visualization tasks, including modern 4D CT scanning methods, represent data natively as INRs. While INRs are prized for being more memory-efficient than traditional data stored on a lattice, many visualization tasks still require discretization to a regular grid. We present PruningAMR, an algorithm that builds a mesh with resolution adapted to geometric features encoded by the INR. To identify these geometric features, we use an interpolative decomposition pruning method on the weight matrices of the INR. The resulting pruned network is used to guide adaptive mesh refinement, enabling automatic mesh generation tailored to the underlying resolution of the function. Starting from a pre-trained INR--without access to its training data--we produce a variable resolution visualization with substantial memory savings.