The dataset is composed of speckle contrast images of microfluidic channels with different flows and different exposure times. Speckle contrast images are provided for two optical magnification under each experimental condition. The dataset has been used to train a convolutionnal neural network for inference of blood flow maps using multiple exposure speckle contrast images as inputs. ( Further details on experimental methods and contents of the dataset are provided in the readme.txt file)
ImageJ, 1.54d
Micro-Manager, 1.43