Abstract:Recent advances in computer vision have led to significant progress in the generation of realistic image data, with denoising diffusion probabilistic models proving to be a particularly effective method. In this study, we demonstrate that diffusion models can effectively generate fully-annotated microscopy image data sets through an unsupervised and intuitive approach, using rough sketches of desired structures as the starting point. The proposed pipeline helps to reduce the reliance on manual annotations when training deep learning-based segmentation approaches and enables the segmentation of diverse datasets without the need for human annotations. We demonstrate that segmentation models trained with a small set of synthetic image data reach accuracy levels comparable to those of generalist models trained with a large and diverse collection of manually annotated image data, thereby offering a streamlined and specialized application of segmentation models. Modern generative techniques have unlocked the potential to create realistic image data of high quality, prompting the possibility of substituting real image data in segmentation training workflows. Our study highlights the capacity of denoising diffusion probabilistic models to generate high-quality microscopy image data. With adjustments to the generation process, these models can produce realistic fully-annotated image datasets through an intuitive and unsupervised approach. The parameters of the generative pipeline undergo optimization through various evaluations, resulting in synthetic image data that exhibits high PSNR scores. Our practical experiments encompass multiple scenarios, including manual annotations, initial segmentations, and simulations as starting points, demonstrating the versatility of our approach. Importantly, we compare the performance of segmentation models trained on a limited set of synthetic image data with those trained on a vast and diverse collection of manually annotated data, demonstrating the potential of our pipeline to alleviate the reliance on extensive manually annotated datasets. Our approach lays the groundwork for similar applications, thereby promoting the much-needed availability of publicly accessible fully-annotated image datasets and advancing the goal of annotation-free segmentation.

Free annotated data for deep learning in microscopy? A hitchhiker's guide

Roadmap on Deep Learning for Microscopy

A Survey on Applications of Deep Learning in Microscopy Image Analysis

Two-stage Generative Models of Simulating Training Data at the Voxel Level for Large-Scale Microscopy Bioimage Segmentation

Advancing electron microscopy using deep learning

DeepBacs for multi-task bacterial image analysis using open-source deep learning approaches

Microsnoop: A Generalized Tool for Unbiased Representation of Diverse Microscopy Images

The Rise of Data-Driven Microscopy powered by Machine Learning

A deep learning dataset for sample preparation artefacts detection in multispectral high-content microscopy

The rise of data‐driven microscopy powered by machine learning

DeepBacs: Bacterial image analysis using open-source deep learning approaches

Denoising Diffusion Probabilistic Models for Generation of Realistic Fully-Annotated Microscopy Image Data Sets

Motivation for using data-driven algorithms in research: A review of machine learning solutions for image analysis of micrographs in neuroscience

Deep-Learning-Based Image Reconstruction and Enhancement in Optical Microscopy

Fluorescent Neuronal Cells v2: Multi-Task, Multi-Format Annotations for Deep Learning in Microscopy

Deep Learning for Automated Experimentation in Scanning Transmission Electron Microscopy

Unsupervised Deep Learning Methods for Biological Image Reconstruction and Enhancement

Review: Deep Learning in Electron Microscopy

Denoising diffusion probabilistic models for generation of realistic fully-annotated microscopy image datasets

Machine Learning for Microscopy Data Analysis: Toward Real-time Optical and Electrical Characterization of Sub-micron Materials

Biological image analysis using deep learning-based methods: Literature review