CRISP: Correlation-Refined Image Segmentation Process.
Jennifer K BriggsErli JinMatthew J MerrinsRichard K P BenningerPublished in: bioRxiv : the preprint server for biology (2024)
Calcium imaging offers the ability to observe cellular activities in real-time across various contexts. However, the manual outlining of cells in calcium imaging data introduces potential errors. This paper introduces the Correlation-Refined Image Segmentation Process (CRISP), an automated algorithm designed to enhance the accuracy of cell mask refinement and to assist in the identification of cell boundaries. CRISP leverages intracell correlations to refine manually drawn masks and automate the detection of the largest cell area that contains only pixels from within the cell. The algorithm not only enhances the precision of calcium trace data but also improves the reliability of functional network analyses.
Keyphrases
- deep learning
- single cell
- cell therapy
- high resolution
- machine learning
- stem cells
- convolutional neural network
- induced apoptosis
- patient safety
- obstructive sleep apnea
- bone marrow
- mass spectrometry
- signaling pathway
- heavy metals
- mesenchymal stem cells
- quality improvement
- sensitive detection
- adverse drug
- photodynamic therapy