Label-free analysis of brightfield microscope images

Provided is a training method (100) for training a predictive model, PM, (132) for highlighting cellular objects, c-objects, in a test Brightfield, BF, image of a test sample (202) of one or more c-objects acquired using a test microscope (204) having at least a BF acquisition mode, wherein the training method (100) uses as training data training Brightfield images (112, 112') and training fluorescent images (114, 114') of a fluorescently labelled training sample as ground truth which are automatically generated. It further relates to method for determining characterising parameters of a test sample, assays and screening method on a test sample that can be label-free.

Patent Publication Number: EP4060552

Year filing: 2021

Year approval: 2023

Year publication: 2022

Status: Requested

Technology domains: undefined

Validated for IOF-key: Yes

Attributed to: Associatie Universiteit & Hogescholen Antwerpen