ObiWan-Microbi

ObiWan-Microbi facilitates semi-automated microbial image annotation to generate reliable ground-truth data for training and evaluating deep learning instance segmentation models in microfluidic live-cell imaging.

Key Features:

• Integrated Workflow: Combines advanced image-analysis technologies into a workflow tailored for microbial image annotation.
• Ground Truth Generation: Produces reliable ground-truth datasets aimed at supporting deep learning, particularly instance segmentation.
• Semi-automated Annotation: Implements a semi-automated approach to image annotation that integrates automated and manual steps.
• Collaborative Annotation: Integrates collaborative features that allow multiple contributors to refine annotations.

Scientific Applications:

• Microfluidic Live-Cell Imaging: Annotation and curation of microbial images from microfluidic live-cell imaging experiments.
• Deep Learning Segmentation: Generation of ground-truth data for development, training, and evaluation of deep learning instance-segmentation models.

Methodology:

ObiWan-Microbi employs a semi-automated approach to image annotation, integrating collaborative features that allow multiple users to contribute to and refine annotations and to generate reliable ground-truth data for training deep learning instance-segmentation models.