| ... | ... | @@ -3,6 +3,7 @@ title: Inspect |
|
|
|
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
In EpiCure, an event is a point (a cell at a given time point) that has been detected as potentially wrongly segmented/tracked and might need manual correction (suspect) or that corresponds to a cellular event (division, extrusion..).
|
|
|
|
|
|
|
|
Several features allows to detect possible segmentation/tracking errors and to display and navigate through these potential errors.
|
| ... | ... | @@ -30,66 +31,15 @@ You can manually add a division event, or remove one. |
|
|
|
* To add a divison, do <kbd>Control+Shift+Left clikc</kbd> from one daughter cell to the other. EpiCure will detect automatically the most likely mother cell.
|
|
|
|
If suspect events had been found on the two daughter cells or on the mother cell, they will be automatically removed when manually adding the division.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# Frame-based suspects
|
|
|
|
|
|
|
|
Cells can be suspected of being un-correctly segmented based on several static features, if their values fall out of range.
|
|
|
|
When a cell is suspected, a point will be added in the [Suspects](#visualization) layer at the center of the cell in the frame where the suspicion has been raised.
|
|
|
|
|
|
|
|
From a static frame, several features can be measured and evaluated for out of range values:
|
|
|
|
|
|
|
|
* [Cell area](#cell-area)
|
|
|
|
* [Tubeness](#tubeness)
|
|
|
|
* [Solidity](#solidity)
|
|
|
|
* [Intensity](#intensity)
|
|
|
|
|
|
|
|
Click on the desired feature button to launch the feature inspection.
|
|
|
|
|
|
|
|
### Options
|
|
|
|
|
|
|
|
`only current frame`: if it is checked, only the current frame will be inspected for the given feature. Else, the inspection will be done on each frame.
|
|
|
|
|
|
|
|
### Shortcuts
|
|
|
|
|
|
|
|
The active layer must be the `Segmentation` layer for the shortcuts to work.
|
|
|
|
|
|
|
|
### Cell area
|
|
|
|
|
|
|
|
This feature measure the cell area in all frames (or only current frame if the option is checked) and test if it is above/below chosen threshold.
|
|
|
|
Each cell (at each time point) below/above the two parameters around the `< area (pix^2) <` buttons will be flagged as suspect for area.
|
|
|
|
|
|
|
|
The parameters are in pixel^2 unit. Use `0` for the min area parameter to keep all small cells, and a very high value for the max area parameter to keep all large cells.
|
|
|
|
|
|
|
|
|
|
|
|
### Intensity
|
|
|
|
|
|
|
|
This feature consider the ratio of intensity of the inner cell region compared to the cell periphery. Indeed, the peripheric part of the cell containing the junction should be much brighter while inside the cell should not be dark. If the ratio is above the chosen threshold, the cell will be marked as suspicious.
|
|
|
|
|
|
|
|
### Tubeness
|
|
|
|
|
|
|
|
This feature consider the ratio of linearity of the inner cell compared to the cell periphery. Indeed, the peripheric part of the cell containing the junction should be linear while inside the cell it should not be linear at all. The "linearity" is quantified by the tubeness of the intensity image. When the ratio is above the chosen threshold, the cell will be marked as suspicious.
|
|
|
|
|
|
|
|
|
|
|
|
### Solidity
|
|
|
|
|
|
|
|
Looks at the cell shape.
|
|
|
|
Cell as detected as outliers if their solidity is greater or smaller than the vast majority of the cells **in the current frame** following Tukey's outlier scheme:
|
|
|
|
|
|
|
|
```plaintext
|
|
|
|
solidity < Q1 - k*(Q3-Q1)
|
|
|
|
solidity > Q3 + k*(Q3-Q1)
|
|
|
|
```
|
|
|
|
|
|
|
|
`Q1` and `Q3` are the 1st and 3rd quartiles. `k` controls the range to consider a cell as outlier and can be chosen in the interface.
|
|
|
|
|
|
|
|
|
|
|
|
---
|
|
|
|
|
|
|
|
|
|
|
|
# Track-based suspects
|
|
|
|
|
|
|
|
__Flag suspicious segmentation based on track examination.__
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
## Ignoring border cells
|
|
|
|
|
|
|
|
Cells on the border of the tissu or imaged field are often flagged as potential errors as they can leave/enter the imaged region and thus have irregular tracks. To not flag these cells, check the `Ignore cells on` `border (of image)` and/or `tissue boundary` options.
|
| ... | ... | @@ -124,6 +74,7 @@ Click on `Inspect track` to launch the analyse with the selected parameters. It |
|
|
|
|
|
|
|
When the process is finished, the number of `suspects` will be updated in the top of the panel and you can see each event as a point (a cross for suspect, circle for division and triangle for extrusions) in the corresponding position in the movie.
|
|
|
|
|
|
|
|
|
|
|
|
# Visualisation
|
|
|
|
|
|
|
|
:round_pushpin: **Press <kbd>x</kbd> to show/hide the suspect points.**
|
| ... | ... | @@ -146,3 +97,58 @@ To directly navigate between suspects, :round_pushpin: **Press <kbd>Space</kbd> |
|
|
|
If the segmentation/tracking is in fact correct, you can remove the point by clicking on `Exonerate current event`. :round_pushpin: **<kbd>Ctrl+Alt+Right click</kbd> to remove the clicked point from event list**
|
|
|
|
|
|
|
|
---
|
|
|
|
|
|
|
|
|
|
|
|
# Frame-based suspects
|
|
|
|
|
|
|
|
Cells can be suspected of being un-correctly segmented based on several static features, if their values fall out of range.
|
|
|
|
When a cell is suspected, a point will be added in the [Suspects](#visualization) layer at the center of the cell in the frame where the suspicion has been raised.
|
|
|
|
|
|
|
|
From a static frame, several features can be measured and evaluated for out of range values:
|
|
|
|
|
|
|
|
* [Cell area](#cell-area)
|
|
|
|
* [Tubeness](#tubeness)
|
|
|
|
* [Solidity](#solidity)
|
|
|
|
* [Intensity](#intensity)
|
|
|
|
|
|
|
|
Click on the desired feature button to launch the feature inspection.
|
|
|
|
|
|
|
|
### Options
|
|
|
|
|
|
|
|
`only current frame`: if it is checked, only the current frame will be inspected for the given feature. Else, the inspection will be done on each frame.
|
|
|
|
|
|
|
|
### Shortcuts
|
|
|
|
|
|
|
|
The active layer must be the `Segmentation` layer for the shortcuts to work.
|
|
|
|
|
|
|
|
### Cell area
|
|
|
|
|
|
|
|
This feature measure the cell area in all frames (or only current frame if the option is checked) and test if it is above/below chosen threshold.
|
|
|
|
Each cell (at each time point) below/above the two parameters around the `< area (pix^2) <` buttons will be flagged as suspect for area.
|
|
|
|
|
|
|
|
The parameters are in pixel^2 unit. Use `0` for the min area parameter to keep all small cells, and a very high value for the max area parameter to keep all large cells.
|
|
|
|
|
|
|
|
|
|
|
|
### Intensity
|
|
|
|
|
|
|
|
This feature consider the ratio of intensity of the inner cell region compared to the cell periphery. Indeed, the peripheric part of the cell containing the junction should be much brighter while inside the cell should not be dark. If the ratio is above the chosen threshold, the cell will be marked as suspicious.
|
|
|
|
|
|
|
|
### Tubeness
|
|
|
|
|
|
|
|
This feature consider the ratio of linearity of the inner cell compared to the cell periphery. Indeed, the peripheric part of the cell containing the junction should be linear while inside the cell it should not be linear at all. The "linearity" is quantified by the tubeness of the intensity image. When the ratio is above the chosen threshold, the cell will be marked as suspicious.
|
|
|
|
|
|
|
|
|
|
|
|
### Solidity
|
|
|
|
|
|
|
|
Looks at the cell shape.
|
|
|
|
Cell as detected as outliers if their solidity is greater or smaller than the vast majority of the cells **in the current frame** following Tukey's outlier scheme:
|
|
|
|
|
|
|
|
```plaintext
|
|
|
|
solidity < Q1 - k*(Q3-Q1)
|
|
|
|
solidity > Q3 + k*(Q3-Q1)
|
|
|
|
```
|
|
|
|
|
|
|
|
`Q1` and `Q3` are the 1st and 3rd quartiles. `k` controls the range to consider a cell as outlier and can be chosen in the interface.
|
|
|
|
|
|
|
|
|
|
|
|
--- |
