Cell Image Analysis Techniques

This paper introduces a method using non-local correlation in the embedding space to improve cell instance segmentation in EM images, reducing over- and under-segmentation issues.

BigNeuron benchmarks automatic neuron reconstruction algorithms using gold standard annotations and image quality metrics, showing that consensus methods outperform individual algorithms.

The paper introduces a segmentation pipeline combining neural networks and active meshes to segment and track cells in 3D time-lapse images of organoids and stem cells.

IntestLine maps rolled intestinal tissue images onto a line to improve visualization and analysis of proximal-distal and serosa-luminal features.

This study systematically quantifies cell morphodynamics in C. elegans embryos, revealing high reproducibility and unique dynamics across cells.

VesselExpress automates 3D blood vasculature analysis in LSFM data, enabling rapid and unbiased quantification of vascular parameters across multiple organs.

The paper provides a searchable image resource of Drosophila GAL4-driver expression patterns at single neuron resolution to improve cell type identification and split-GAL4 combination prediction.

The paper develops optimized small CNNs for cell shape detection in confluent tissues, finding they outperform transfer learning in accuracy, speed, and simplicity.

petiteFinder is a deep learning tool that automates Petite colony frequency calculation in baker's yeast by analyzing images of Petri dishes.

MoBIE is a Fiji plugin that enables visualization, sharing, and management of large, multi-modal image data across platforms.

CASSATT processes cycIHC data to align images, segment nuclei, and analyze spatial cell interactions and neighborhoods.

Fast4DReg corrects drift in 3D microscopy datasets using 2D cross-correlations, outperforming existing tools in speed and accuracy.

SUFI automates spectral unmixing of fluorescent multiplex images in mouse and postmortem human brain tissues, improving accuracy over manual methods and existing software.

GliaMorph quantifies Müller glial cell morphology across developmental and disease states, revealing structural changes and subcellular features in zebrafish and mouse models.

The paper introduces an unsupervised method to extract morphological features from 3D EM data, enabling cell type and tissue organization analysis in a segmented worm.

This paper introduces a Transformer-based method for single particle tracking in fluorescence microscopy images, showing improved performance over existing methods.

This study compares linear and SRV metrics for analyzing tumor cell shape heterogeneity, finding SRV provides better separation and low-dimensional representation of cell morphology.

This paper presents a segmentation workflow that enables cell segmentation without structural fluorescent labels, improving performance and robustness across cell lines and fluorescent labels.

The paper introduces three plugins for visualization and quality control in large-scale multiplex tissue analysis to improve accuracy and user trust in automated approaches.

The paper uses Cell Painting images to preselect compounds for drug screening, significantly increasing hit rates by reducing the number of compounds tested.

This paper develops a weakly-supervised method using recurrent neural networks to classify cell-cycle stages in mitosis videos without annotated data, achieving performance comparable to supervised approaches.

The paper introduces MOAProfiler, a deep learning encoder that improves the accuracy of drug mechanism of action classification from Cell Painting images compared to traditional and existing deep learning methods.

This paper introduces an auxiliary learning task using Local Shape Descriptors to improve neuron segmentation in electron microscopy volumes, achieving performance comparable to state-of-the-art methods with greater efficiency.

This paper develops a nonlinear image registration method for live cell movies to analyze spatiotemporal dynamics of diffuse molecular processes, revealing organized Profilin dynamics linked to actin organization.

SpheroScan is a deep learning tool for automated spheroid image analysis that improves reproducibility and throughput in 3D spheroid assays.

The paper introduces an ImageJ-based tool for 3D registration between microscopic images, enabling alignment and reconstruction of 3D rotated tissues like mouse embryos and organoids.

FriendlyClearMap simplifies and extends the ClearMap pipeline for mouse brain analysis, enabling efficient and scalable neuroanatomical studies.

The paper introduces TACI, an ImageJ plugin for analyzing 3D calcium imaging data, improving reproducibility and reducing human bias without requiring coding.

The paper evaluates cell segmentation methods using automated metrics instead of human reference segmentations, finding two deep learning methods perform best with improved results via postprocessing.

This paper develops a scalable 3D imaging protocol to measure DNA damage response in ovarian cancer organoids using γH2AX staining and shows its effectiveness with Etoposide treatment.