Knowledge Base

Troubleshooting

The Paw Prints Look Faint, or Irregular

This may have multiple causes: (a) the camera gain is too low, (b) the walkway lights are too dim, or (c) the green intensity threshold is too high.

1. In the classification screen, click the Adjust the Green Intensity Threshold button. Deselect Show Filtered Out Signals. When you move the slider to the right, the prints should shrink and then disappear. When you move the slider to the left, more and more of the prints should appear in green. Change this setting until the prints look good. However this change only affects that run. If that solves the issue, in the Detection Settings, make the same change to the green intensity threshold there to apply to all future runs.
2. If the previous step does not help, the issue may lie in the camera gain or the lights. Go to the Detection Settings (Setup > Detection Settings). Set the Green Intensity Threshold fairly low, for example 0.05. With an animal running across the walkway, look at the value in the middle for Maximum Green Intensity. During a run, this value should be around 200–205. If the animal stops, or walks more times in the same place, the value increases. Click the Reset Prints button to clear that value and try again. If the Maximum Green Intensity is much lower than about 200, raise the Camera Gain. If you raise the Camera Gain to 20 or higher and the maximum green intensity is still too low, also raise the Green Walkway Light setting (only a little bit at a time). Keep clicking Reset Prints while the animal runs back and forth to get new measures. Avoid having the camera gain above 20 as it produces more noise.
3. Once the Camera Gain is set, with a low Green Intensity Threshold you will probably also have a lot of other noise in the image. Now start adjusting that threshold value. You want to get nice clean images of the prints but no green noise anywhere else. If you raise the Green Intensity Threshold too high, you'll start losing the prints themselves. Play with this setting till the prints look good.

A Paw Print Is Not Recognized by the Automatic Footprint Classification

This may happen when a paw print has consistently a low value of Max Intensity, and this value is lower than the minimum set in the Classification Settings.

Choose Setup > Classification Settings. Next to Only Assign a Label..., reduce the value of minimum acceptable Max Intensity. Next, repeat automatic classification.

The Automatic Footprint Classification Does Not Seem to Work Well

This may occur in two cases: (a) when prints are too light, because the gain in the detection settings is too low, or the noise threshold is too high, or (b) when there is too much noise that confuses the system, because the gain is too high or the noise threshold too low.

Especially in the case where there is too much noise, you can adjust the noise threshold in the classification. Click the button Adjust the Green Intensity Threshold and use the slider to raise or lower the threshold. A lower threshold will show more of the prints and noise, while a higher threshold will remove more of them. Your goal is to have only the prints left, as complete as possible, but no noise. The closer you get to that situation, the better a job auto classification will do.

If you consistently need to adjust the noise threshold, it's best to do so in the Detection Settings. That is also the only place you can adjust the camera Gain (sensitivity). The Gain can only be adjusted for future trials, not for those which have already been acquired.

Paw Prints Look Incomplete in the Interactive Footprint Measurements Screen

When the footprint is shown in the Interactive Footprint Measurements screen, it is enlarged by means of smoothing and anti-aliasing algorithms. Pixels that are outside the border of the footprint rectangle, which contain no data by definition, are "translated" to blue pixels. This may result in incomplete-looking images, especially if the image resolution is low. To reduce this effect, place the camera closer to the walkway.

Tip: In the Interactive Footprint Measurements screen, right-click the footprint image and select Switch to 'Normal Color' Mode. That way the artifact is less evident.

I Want to Include Non-Compliant Runs in My Data Set

In the lower-left corner, under Data Segmentation Profiles, double-click All Data. In the window that opens, make sure that the two check boxes at the top are not selected. Then click OK and run analysis.

How Can I Separate Two Paw Prints That Are Near Each Other?

When two paw prints are very near each other, drawing a box around one of the prints is likely to include parts of the other print.

1. Draw a box that only covers (part of) the first paw print, and select Add Print. If a box is already defined around both paw prints, remove that box (right-click and select Remove Print).
2. Draw more boxes for the remaining parts of the paw print, but try not to include the second paw print. Keep selecting Add Print. Each box is now labeled as UN.
3. Draw a larger box around all the boxes and select the paw print you want.
4. Repeat the steps for the other paw print.

The Automatic Detection of Paw Prints Does Not Work Anymore

This could be caused by the glass plate no longer being in its place. In that case illumination is no longer uniform; that is, paw prints are found only on the left or right parts of the catwalk. Check that the glass plate has not been pulled or pushed during transport of the system, and that both the left and right sides of the walkway are equally visible. If you move the system to another location, always do this by pushing/pulling the whole CatWalk stand, not the glass plate.

Data Analysis and Export

The Group Means Do Not Correspond to What I Calculate Using the Exported Data

This occurs because of the way the group means are calculated. In CatWalk XT, the group mean of a variable is calculated by taking the means of the trials that correspond to that group. So if a group contains 10 trials, the mean is calculated over the 10 trial averages, irrespective of the "weight" of each trial, that is, of how many runs each trial contains. If you consider the single runs, the overall mean value for that group will differ from that given by CatWalk XT.

The Circular Charts Look Distorted When Exported

When being exported to image files, the circular charts may look stretched or squeezed. To solve this, click the Chart Settings button and select the same value of Chart Width, Chart Height, and Chart Size (the latter under Circular Charts). Then export the charts again.