Data format
Subfolders
Subfolders
Each dataset contains four sub-folders which contain the FOV images (.png), annotation coordinates (.csv), mask images (.png), and visualization images (.png). All images are at 0.2 microns-per-pixel. All coordinate values in tables are pixel units at 0.2 microns-per-pixel.
Each dataset contains four sub-folders which contain the FOV images (.png), annotation coordinates (.csv), mask images (.png), and visualization images (.png). All images are at 0.2 microns-per-pixel. All coordinate values in tables are pixel units at 0.2 microns-per-pixel.
rgb
rgb
csv
csv
mask
mask
visualization
visualization
Naming convention
Naming convention
File names encode most relevant information. There may be a prefix encoding participant and dataset.
File names encode most relevant information. There may be a prefix encoding participant and dataset.
Coordinate files (.csv)
Coordinate files (.csv)
Each contains nuclei coordinates for the corresponding image. Additionally, a file called all_fov_locations.csv contains the intended annotation region. This accounts for the fact that nuclei may extend beyond the FOV boundary.
Each contains nuclei coordinates for the corresponding image. Additionally, a file called all_fov_locations.csv contains the intended annotation region. This accounts for the fact that nuclei may extend beyond the FOV boundary.
> raw_classification: raw class (13 total)
> raw_classification: raw class (13 total)
> main_classification: nucleus class (7 total)
> main_classification: nucleus class (7 total)
> super_classification: nucleus superclass (4 total)
> super_classification: nucleus superclass (4 total)
> type: rectangles v.s. polylines
> type: rectangles v.s. polylines
> xmin / ymin / xmax / ymax: extent of the nucleus
> xmin / ymin / xmax / ymax: extent of the nucleus
> coords_x / coords_y: comma-separated boundary
> coords_x / coords_y: comma-separated boundary
Mask format
Mask format
Mask images are provided for convenient use with machine learning analyses. The first channel in each mask image encodes the class labels found here. The product of the second and third channels encode the unique instance label for each nucleus. The fov area (gray) is included in the class table and first channel of the mask. This file contains the nucleus label encoding, including a special 'fov' code encoding the intended annotation region.
Mask images are provided for convenient use with machine learning analyses. The first channel in each mask image encodes the class labels found here. The product of the second and third channels encode the unique instance label for each nucleus. The fov area (gray) is included in the class table and first channel of the mask. This file contains the nucleus label encoding, including a special 'fov' code encoding the intended annotation region.
Note: unlike the csv files, masks do not distinguish between bounding boxes and segmentations.
Note: unlike the csv files, masks do not distinguish between bounding boxes and segmentations.
1st channel
1st channel
2nd x 3rd channels
2nd x 3rd channels
Clustering results
Clustering results
Anchors obtained from clustering are in files named like: v3.1_final_anchors_E_Ps_AreTruth.csv. In this case, this file includes anchors that were collectively judged by pathologists to be "real" nuclei. Some relevant columns are highlighted here. Please contact us with any questions.
Anchors obtained from clustering are in files named like: v3.1_final_anchors_E_Ps_AreTruth.csv. In this case, this file includes anchors that were collectively judged by pathologists to be "real" nuclei. Some relevant columns are highlighted here. Please contact us with any questions.
> xmin_relative, ymin_relative, .. : anchor extent
> xmin_relative, ymin_relative, .. : anchor extent
> MV_[...] : majority voting results
> MV_[...] : majority voting results
> EM_[...] : expectation-maximization results
> EM_[...] : expectation-maximization results
> NP.1, NP.2, SP.1, .. : matched annotations
> NP.1, NP.2, SP.1, .. : matched annotations
> manual_coords_x/y : manual boundary by one SP
> manual_coords_x/y : manual boundary by one SP
> algorithmic_coords_x/y : suggested boundary
> algorithmic_coords_x/y : suggested boundary