Cell Image Analysis Techniques

Anatolution provides an open-source platform for generating consensus-validated training data for cell-level morphometry in histology.

This paper introduces a robotic system for high-throughput single-cell manipulation in human brain organoids using vision guidance.

The paper introduces a Python module for accessing TrypTag data on protein subcellular localization in Trypanosoma brucei and demonstrates its use in analyzing age-based organelle composition differences.

The paper introduces TGlow, a high-content imaging platform for single-cell phenotyping of human lymphocytes, revealing distinct morphological states and drug responses.

This paper harmonizes 2D and 3D immunofluorescent cell shape datasets using generative AI to create synthetic cells for machine learning and in silico studies.

HFinder uses deep learning to automate and quantify analysis of plant-microbe interfaces in confocal images, enabling detailed study of subcellular interactions.

This paper develops a deep learning tool for automating cell line development, achieving a 36-day workflow for generating high-titer antibody-producing cell lines.

BlueNuclei automates the identification and classification of live and dead transfected neurons using interpretable features, achieving near-human accuracy with faster processing than deep learning alternatives.

SpatialDINO enables automated segmentation and tracking of crowded cellular objects using self-supervised 3D vision transformers without manual annotations or retraining.

This study compares supervised machine learning and expert assessment in quantifying ultrastructural changes in macrophages, finding sML provides more objective and sensitive analysis of subtle differences.

CellFluxV2 is an image-generative model that predicts cell morphology changes under perturbations, achieving higher image fidelity than baselines and enabling batch-effect correction and cell-state interpolation.

This paper introduces a deep learning model to identify virus-infected cells from light microscopy images without using labels or stains, achieving up to 96% classification accuracy.

This paper introduces an automated image-based cell sorting method that enriches activated T cells from blood with high throughput and label-free morphology-based classification.

This paper introduces a dataset of high-resolution fluorescence images of astrocytic gap junctions to support machine learning development for automated analysis.

CLaSSiNet classifies nanoscale membrane skeleton networks with high resolution, revealing cell-type-specific organizational patterns and mechanical coupling principles.

SpineDL is a deep learning tool for segmenting neurons and anatomical structures in spinal cord immunofluorescence images, achieving expert-level performance.

The paper introduces a bioimaging data processing pipeline that improves data transformation efficiency and accelerates downstream tasks by optimizing I/O throughput and reducing memory overhead.

The paper develops an image-based pipeline to analyze autophagy changes in ZIKV-infected cells and trains a deep learning classifier to identify infected cells using autophagy reporters.

This paper introduces a deep learning framework for classifying calcium imaging dynamics using quantile-based time-series representations, showing strong performance across different preprocessing strategies and noise conditions.

The paper develops an AI algorithm to classify Gram stain images into six categories, achieving 99.9% accuracy using machine learning after initial rule-based methods.

The paper develops a universal imaging phenome using SAM features and SPOT tool, showing it outperforms deep learning models in capturing dynamic cell behaviors.

This study uses image-based morphological profiling to detect radiation exposure signatures in fibroblasts, showing morphological changes as early as 4 hours post-exposure.

SPACEMAP is a machine learning framework for multiplex imaging analysis that resolves phenotyping discordance by integrating multiple classification methods.

This paper develops a high-fidelity 3D reconstruction pipeline for murine lymph nodes, enabling accurate modeling of drug transport and immune cell dynamics.

CLEAR-AF improves autofluorescence subtraction in multiplexed tissue imaging by using polar transformation and Gaussian Mixture Models, enhancing specificity, sensitivity, and reproducibility.

MExConn is a multi-expert U-Net framework for segmenting multiple organelles in EM images, outperforming baselines and offering mechanistic interpretability.

CochleaNet is a deep learning framework for analyzing cochlea imaging data, enabling segmentation and analysis of cochlear structures and gene therapy products.

MAPL3 enables whole-brain projectome mapping at single-axon resolution using deep learning and light sheet microscopy, outperforming existing methods across diverse datasets.

The paper profiles the morphological effects of 2,565 human and 1,900 mouse miRNAs on 24 million cells across five cell lines, finding that 9% and 15% of miRNA mimics significantly alter cell morphology.

MitoEM 2.0 provides a benchmark for 3D mitochondria segmentation in EM images with diverse datasets and standardized labels for model evaluation and comparison.