VesiScope: A stand-alone tool for the automatic detection, quantification and size analysis of GUVs from phase contrast microscopy images

Abstract

Fluorescence labeling is commonly used to enhance microscopy image contrast and facilitate automated Giant Unilamellar Vesicles (GUVs) quantification. Nonetheless, fluorescent probes may introduce unaccounted variations compared to the intended pristine formulation. Alternatively, phase-contrast microscopy enables label-free GUVs observation, but the lack of reliable automated detection methods limits high throughput analyses. To address this gap, we developed a framework based on YOLOv11 model, a state-of-art lightweight convolutional neural network for object detection. Trained on 4666 manually annotated GUVs from 164 phase-contrast microscope images, the model enables robust detection under low contrast-to-noise ratio and heterogeneous illumination. The trained model is implemented in a stand-alone desktop application supporting multiple analysis in a single run, pixel-to-micrometer calibration, and automated size distribution analysis. Retraining scripts compatible with standard hardware are also provided to enhance adaptability. VesiScope achieves high detection accuracy and competitive cross-microscope generalization enabling fast and reliable analysis of unaltered GUVs, supporting experimental reproducibility.