Towards 1000-fold Electron Microscopy Image Compression for Connectomics via VQ-VAE with Transformer Prior
This addresses storage and transfer bottlenecks in connectomics, offering a domain-specific solution for handling large-scale EM data.
The paper tackles the challenge of compressing petascale electron microscopy datasets by introducing a VQ-VAE-based framework that achieves up to 1024x compression, enabling pay-as-you-decode usage and selective high-resolution reconstruction from compressed latents.
Petascale electron microscopy (EM) datasets push storage, transfer, and downstream analysis toward their current limits. We present a vector-quantized variational autoencoder-based (VQ-VAE) compression framework for EM that spans 16x to 1024x and enables pay-as-you-decode usage: top-only decoding for extreme compression, with an optional Transformer prior that predicts bottom tokens (without changing the compression ratio) to restore texture via feature-wise linear modulation (FiLM) and concatenation; we further introduce an ROI-driven workflow that performs selective high-resolution reconstruction from 1024x-compressed latents only where needed.