AboutDownloadsDocumentsForumsIssuesNews
This package contains the dataset for the study. Each non-label image is a grayscale 2560*1920 pixel 8bit acquired with a phase-contrast microscope. Images in the training and validation set contains only a single genus of bacteria (i.e. they are single-class) and they are named accordingly to reflect the class, e.g. "Lb" contains Lactobacillus. Each image has a corresponding labelled mask which has the same name as its corresponding image but with the suffix "_P.png". In these masks each pixel containing background is labelled '0' and each pixel belonging to a cell contains the number 1, 2, 3, 4 or 5 depending on the class it belongs to. The list of classes is contained in the "Genera.txt" file. Two sets of backgrounds are also provided in the dataset. For instructions on how to use these, refer to the publication. The training set is comprised of the files: Training1, Training2, Training3, Training4, Training5, Training6 and Training7. The validation set is comprised of the files: Validation1, Validation 2. The labels contain labelled masks for both the training set and validation set and is comprised of the files: Labels1. The testsets are comprised of two sets, one with a mix of bacteria together and one with mixed Lactococcus bacteria with milk protein. Each image has a corresponding labelled mask. The testset is comprised of the files: Testsets1. The backgrounds are comprised of two sets, one with low or no objects in them and one with milk protein. The backgrounds are comprised of the files: Backgrounds1, Backgrounds2.
Downloads instructions and details
This project has no publicly available downloads.
To access any private downloads that may exist, please log in.
Feedback