{\mu}Split: efficient image decomposition for microscopy data

TL;DR

{}Split introduces a memory-efficient deep learning framework with lateral contextualization for improved image decomposition in microscopy, enabling larger context use, deeper models, and reduced artifacts, outperforming existing methods.

Contribution

The paper proposes }Split, a novel meta-architecture with lateral contextualization that enhances deep models for microscopy image decomposition while reducing memory usage and artifacts.

Findings

Achieves an average of 2.25 dB PSNR improvement over baselines.

Requires significantly less GPU memory than traditional methods.

Effectively reduces tiling artifacts in hierarchical models.

Abstract

We present {\mu}Split, a dedicated approach for trained image decomposition in the context of fluorescence microscopy images. We find that best results using regular deep architectures are achieved when large image patches are used during training, making memory consumption the limiting factor to further improving performance. We therefore introduce lateral contextualization (LC), a novel meta-architecture that enables the memory efficient incorporation of large image-context, which we observe is a key ingredient to solving the image decomposition task at hand. We integrate LC with U-Nets, Hierarchical AEs, and Hierarchical VAEs, for which we formulate a modified ELBO loss. Additionally, LC enables training deeper hierarchical models than otherwise possible and, interestingly, helps to reduce tiling artefacts that are inherently impossible to avoid when using tiled VAE predictions. We apply {\mu}Split to five decomposition tasks, one on a synthetic dataset, four others derived from real microscopy data. Our method consistently achieves best results (average improvements to the best baseline of 2.25 dB PSNR), while simultaneously requiring considerably less GPU memory. Our code and datasets can be found at https://github.com/juglab/uSplit.