test-data/cellfinder/edge_cells_brain/info.txt

# Test data info (for internal use)

All this data was generated using the main brach (commit a28305618a228fc5175e05f72b555d7b29073bc5) before the pytorch detection PR.

The peak_enhanced and 2d_filter images in this folder were derived from the images in the test
structure_split_test/signal directory. They were generated by passing them
through the 2d filters (median, gaussian, laplacian) using values of
`soma_diameter=16, log_sigma_size=0.2, n_sds_above_mean_thresh=10`. For peak_enhanced, the input data type was uint32 and filtering data type float64.
For 2d_filter it was uint16 and float32, respectively.

The tiles directory contains the output from `TileProcessor.get_tile_mask`, indicating
for each tile whether the tile is in the brain or not.

3d_filter is the output generated by the volume filter after 2d/3d filtering. Using
`voxel_sizes=(5, 2, 2), soma_diameter=16, max_cluster_size=100000, ball_xy_size=6,
ball_z_size=15, ball_overlap_fraction=0.6, soma_spread_factor=1.4, log_sigma_size=0.2,
n_sds_above_mean_thresh=10`.

detected_cells.xml is the cells that were detected by running through the whole process.