The Data Sets - Visual Cortical Layer 6 Neuron
Folder Contents:
This is 1 of the 6 data sets used for DIADEM.
This data set consists of 25 separated dendritic and axonal subtrees from 1 primary visual cortical neuron. The material for this neuron was sectioned into 5 physical slices, 4 of which are used in this data set. Each of the 4 sections consists of numerous image stack tiles. Each image stack tile captures some portion of 1 or more of the 25 tree samples.
During the DIADEM Grand Challenge Competition: This data set was not provided in the Training or Qualifier Rounds and thus served as a ‘wildcard’ during the Final Round to test how algorithms might do on data that had not been used during algorithm development.
Experimental Procedures:
Data set owner: J. A. Hirsch, Department of Biological Sciences, University of Southern California
Species: Cat
Nervous System Region: Primary Visual Cortex Layer 6
Fiber type: Dendrites and Axons
Labeling Method: Biocytin
Image Acquisition Method: Transmitted Light Brightfield
Tracing Method: Neurolucida (Williston, VT)
Objective Lens: 63x
See Martinez et al., 2005, Nat Neurosci. Vol. 8(3) for general information related to the data
Z distance between successive images within an image stack: 2.93 pixels
Data Set Instructions:
1. Download and extract 'Visual Cortical Layer 6 Neuron Part_1.rar' through 'Visual Cortical Layer 6 Neuron Part_10.rar'. Note that these RAR files are independently compressed and must be extracted individually.
2. All RAR files contain either the folder ‘Gold Standard Reconstructions’, ‘Image Stacks- 16 bit Gray’, and/or ‘Image Stacks- 24 bit RGB’, each containing a portion of the data from these folders. Extracted RAR files should be merged to have the following folder hierarchy: Create the parent folder 'Visual Cortical Layer 6 Neuron’; copy all data into the subfolders ‘Gold Standard Reconstructions’, ‘Image Stacks- 16 bit Gray’, and ‘Image Stacks- 24 bit RGB’.
3. 2 image formats have been made available. ‘Image Stacks- 16 bit Gray’ preserves all of the information from the original data but looks grainy in the most recent Neuromantic version (1.7.5). ‘Image Stacks- 24 bit RGB’ preserves only 8 of the original 12 bit information but looks better in Neuromantic. Both image formats differ only in bit encoding so all information provided in this readme applies to both.
4. Image stack folders will contain ‘Section 1’ through ‘Section 4’, which contain subfolders for each of their corresponding image stack folders. Image stack folders are named by Column and Row, which references their tiling position with respect to the other image stack tiles within the section. For example, image stack ‘C01R2’ is located in the leftmost column and just below ‘C01R1’ (see instruction 15 below for explanation of image stack folders with ‘Processed’ in their name that are found in the '24 bit RGB' folder'; see 'Individual Stack Information' section for explanation of why some column and row numbers are not provided).
5. Neighboring tiles are aligned such that they are offset in both X and Y. For example, image stack tile ‘C02R2’ will be to the right and may also be a bit below ‘C02R3’ in Y. Thus, the terms ‘Columns’ and ‘Rows’ are used inexactly.
6. Image stack tiles can overlap with each other and can therefore contain redundant structural information.
7. All image stack tiles within a section are already aligned in Z and so require no Z translation.
8. All gold standard reconstructions will automatically align with image stack tile C01R1 from their respective sections when loaded into software that uses pixel-based coordinates, such as ImageJ or Neuromantic. This remains true for sections where C01R1 is not provided such as for Sections 2, 3 & 4 (i.e. SWC files will align where C1R1 would be located).
9. Each tree sample has starting coordinates provided below.
10. Image stacks may contain overlapping structures that are separate from the given tree sample.
11. The starting coordinates given below clearly mark the root of each tree sample to reconstruct within its given image stack section.
12. Special note for this data set: This data set may contain some reconstructions that do not bifurcate; consisting of only a root and a single termination node.
13. Special note for this data set: All gold standard reconstructions of tree samples were manually reconstructed in 3D, but the Z coordinates in between bifurcation nodes, terminations, and root were smoothed (i.e. interpolated) prior to submission. Therefore, automated reconstructions should use linear connections between all branch segments (i.e. between bifurcations, and terminations, and root) in order to compare to the gold standard reconstructions when using the DIADEM metric.
14. Special note for this data set: Image ‘001.TIF’ of ‘Section 3’ image stack ‘C09R1’ appears to be incorrect as it is inconsistent with the rest of the images. The incorrect image was preserved within the image stack as it was originally. It is suggested that you remove this image from the stack and replace it, possibly with a copy of the image below it, if you are concerned that the inconsistent image will affect performance.
15. The image stacks in the 'Images Stacks- 24 bit RGB' folder are noticeably dark and had to be adjusted in brightness and gain to be utilized during data set quality checking. These adjusted image stacks are provided in the ‘Processed’ folders. 16 bit gray image stacks did not require any such processing.
16. Special note for this data set: External branches (i.e. branches that end in a termination) with path lengths of less than 50 pixels will be ignored when using the DIADEM metric. Thus, they can be traced or ignored without affecting the DIADEM metric score, provided that the traced branches remain less than 50 pixels in length.
17. Special note for this data set: The same systematic method used in the ‘Hippocampal CA3 Interneuron’ data set to decide where structures ended when they leave a given section was used for this data set and should be followed when attempting to achieve results similar to the gold standard reconstructions (see instruction 12 in the 'Hippocampal CA3 Interneuron Readme' README).
Gold standard reconstructions ‘S1_01.swc’ (pink) and ‘S1_02.swc’ (blue) are superimposed over image stacks ‘C01R1’ and ‘C01R2’ from 'Section 1'. Red arrows point to the root of each reconstruction.
Individual Stack Information:
All image stacks are 1024x1024 pixels in XY. Both 16 bit per pixel grayscale and 24 bit per pixel RGB are available.
Note that some columns and rows are not provided since the corresponding image stacks did not contain any of the tree samples used. Column and row names are preserved from the original mosaic from which they came, which was in a custom matrix format (some columns are below other columns in Y). Their spatial relationship is preserved by the translation values below so understanding the details regarding this naming convention is unnecessary.
The pixel-based values below describe how far to move image stacks when they load at (X,Y) = (0,0) at the upper left corner of the image stack. Note that Neuromantic (as of version 1.7.5) can load multiple and move separate image stack tiles (see Instruction 12 in ‘General Readme’):
Section 1 (Number of images per stack: 172)
C01R1- Translation (X,Y): (0,0)
C01R2- Translation (X,Y): (-78,967)
C01R3- Translation (X,Y): (-154,1910)
C02R2- Translation (X,Y): (832,1033)
C02R3- Translation (X,Y): (755,1986)
C03R2- Translation (X,Y): (1753, 1113)
Section 2 (Number of images per stack: 188)
C01R2- Translation (X,Y): (-78,962)
C01R3- Translation (X,Y): (-155,1916)
C02R3- Translation (X,Y): (750,1987)
C03R3- Translation (X,Y): (1645,2059)
C04R1- Translation (X,Y): (-217,2796)
C04R2- Translation (X,Y): (-285,3660)
C05R1- Translation (X,Y): (683,2867)
C05R2- Translation (X,Y): (615,3733)
C06R1- Translation (X,Y): (1575,2939)
C07R1- Translation (X,Y): (2461,3000)
Section 3 (Number of images per stack: 171)
C01R3- Translation (X,Y): (-151,1884)
C03R3- Translation (X,Y): (-2192,4410)
C04R1- Translation (X,Y): (-1128,2712)
C04R3- Translation (X,Y): (-1268,4486)
C05R1- Translation (X,Y): (-222,2783)
C05R2- Translation (X,Y): (-291,3669)
C05R3- Translation (X,Y): (-361,4557)
C06R3- Translation (X,Y): (539,4630)
C07R1- Translation (X,Y): (-3193,5250)
C08R1- Translation (X,Y): (-2255,5322)
C08R2- Translation (X,Y): (-2331,6275)
C08R3- Translation (X,Y): (-2407,7199)
C09R1- Translation (X,Y): (-1338,5397)
C09R2- Translation (X,Y): (-1414,6354)
C10R1- Translation (X,Y): (-434,5468)
C10R2- Translation (X,Y): (-512,6426)
C11R1- Translation (X,Y): (460,5542)
C11R2- Translation (X,Y): (383,6495)
C12R1- Translation (X,Y): (1354,5612)
C12R2- Translation (X,Y): (1277,6567)
C13R1- Translation (X,Y): (2287,5682)
C13R2- Translation (X,Y): (2207,6644)
Section 4 (Number of images per stack: 176)
C01R2- Translation (X,Y): (-75,934)
C01R3- Translation (X,Y): (-150,1874)
C03R3- Translation (X,Y): (-1268,4479)
C04R3- Translation (X,Y): (-361,4550)
C10R1- Translation (X,Y): (-434,5468)
C10R3- Translation (X,Y): (-580,7335)
C11R1- Translation (X,Y): (458,5538)
C11R2- Translation (X,Y): (379,6497)
C11R3- Translation (X,Y): (308,7406)
C12R1- Translation (X,Y): (1377,5613)
C12R2- Translation (X,Y): (1300,6571)
C13R2- Translation (X,Y): (2225,6645)
C14R2- Translation (X,Y): (3140,6719)
C14R3- Translation (X,Y): (3066,7631)
C15R2- Translation (X,Y): (4047,6792)
Gold Standard Reconstruction Starting Coordinates (X and Y in pixels; Z in image sequence number, where top image Z = 0).
Section_1
S1_01 (X, Y, Z): (279, 1124, 45)
S1_02 (X, Y, Z): (439, 1391, 151)
S1_03 (X, Y, Z): (805, 1004, 153)
S1_04 (X, Y, Z): (745, 1616, 153)
Section_2
S2_01 (X, Y, Z): (745, 3423, 163)
S2_02 (X, Y, Z): (1143, 3766, 158)
S2_03 (X, Y, Z): (520, 3834, 159)
Section_3
S3_01 (X, Y, Z): (-371, 2734, 155)
S3_02 (X, Y, Z): (10, 4050, 154)
S3_03 (X, Y, Z): (-1808, 6222, 30)
S3_04 (X, Y, Z): (-1889, 6006 25)
S3_05 (X, Y, Z): (-867, 5155, 27)
S3_06 (X, Y, Z): (2583, 7015, 162)
S3_07 (X, Y, Z): (502, 4884, 40)
S3_08 (X, Y, Z): (-1660, 4954, 45)
S3_09 (X, Y, Z): (-293, 6210, 31)
Section_4
S4_01 (X, Y, Z): (254, 2729, 19)
S4_02 (X, Y, Z): (1696, 6780, 53)
S4_03 (X, Y, Z): (1170, 6915, 46)
S4_04 (X, Y, Z): (2392, 7209, 43)
S4_05 (X, Y, Z): (1308, 5880, 163)
S4_06 (X, Y, Z): (2855, 7335, 46)
S4_07 (X, Y, Z): (3359, 7477, 104)
S4_08 (X, Y, Z): (3425, 7460, 131)
S4_09 (X, Y, Z): (3412, 7510, 105)