Hello @Luki132,

I tried to replicate your issue with multiple export formats but failed to do so.

Can you tell me which export format you selected and maybe provide an example point cloud and label?

Also providing the full config.ini and _classes.json would help.

I'm having the same issue described above. Every time I relaunch the app, the labels shift in height.

Laptop:
OS - Windows 11
Python version - 3.9.0
Installation - manual/git

Error:
I also receive the following errors when launching the app (I thought it might be related).
"Could not create pixmap from C:UsersxperimentallabelCloudlabelCloudresourcesicons\cube-outline.svg"
"Could not create pixmap from C:UsersxperimentallabelCloudlabelCloudresourcesicons\cube-outline_white.svg"
The path here is incorrect as it is missing the "" symbols. It should be: "C:\Users\Experimental\labelCloud\labelCloud\resources\icons\cube-outline.svg"
"C:\Users\Experimental\labelCloud\labelCloud\resources\icons\cube-outline_white.svg"

config.ini:
[FILE]
; source of point clouds
pointcloud_folder = pointclouds/
; sink for label files
label_folder = labels/
; definition of classes and export format
class_definitions = labels/_classes.json
; only for kitti = calibration file for each point cloud
calib_folder = calib/
; sink for segmentation files (*.bin point clouds) [optional]
segmentation_folder = labels/segmentation/
; 2d image folder [optional]
image_folder = images/

[POINTCLOUD]
; drawing size for points in point cloud
point_size = 4.0
; point color for colorless point clouds (r,g,b)
colorless_color = 0.9, 0.9, 0.9
; colerize colorless point clouds by height value [optional]
colorless_colorize = True
; standard step for point cloud translation (for mouse move)
std_translation = 0.03
; standard step for zooming (for scrolling)
std_zoom = 0.0025
; blend the color with segmentation labels [optional]
color_with_label = True
; mix ratio between label colors and rgb colors [optional]
label_color_mix_ratio = 0.3

[LABEL]
; number of decimal places for exporting the bounding box parameter.
export_precision = 8
; default length of the bounding box (for picking mode)
std_boundingbox_length = 0.75
; default width of the bounding box (for picking mode)
std_boundingbox_width = 0.55
; default height of the bounding box (for picking mode)
std_boundingbox_height = 0.15
; standard step for translating the bounding box with button or key (in meter)
std_translation = 0.03
; standard step for rotating the bounding box with button or key (in degree)
std_rotation = 0.5
; standard step for scaling the bounding box with button
std_scaling = 0.03
; minimum value for the length, width and height of a bounding box
min_boundingbox_dimension = 0.01
; propagate labels to next point cloud if it has no labels yet
propagate_labels = False

[USER_INTERFACE]
; only allow z-rotation of bounding boxes. set false to also label x- & y-rotation
z_rotation_only = False
; visualizes the pointcloud floor (x-y-plane) as a grid
show_floor = False
; visualizes the object's orientation with an arrow
show_orientation = True
; background color of the point cloud viewer (rgb)
background_color = 100, 100, 100
; number of decimal places shown for the parameters of the active bounding box
viewing_precision = 2
; near and far clipping plane for opengl (where objects are visible, in meter)
near_plane = 0.1
far_plane = 300
; keep last perspective between point clouds [optional]
keep_perspective = False
; show button to visualize related images in a separate window [optional]
show_2d_image = True
; delete the bounding box after assigning the label to the points [optional]
delete_box_after_assign = False

_classes.json:
{
"classes": [
{
"name": "Car",
"id": 0,
"color": "#0000ff"
}
],
"default": 0,
"type": "object_detection",
"format": "kitti",
"created_with": {
"name": "labelCloud",
"version": "1.1.0"
}
}

Let me know if you need any more information.

@NArcher22 thanks for the data!

Might be Windows-related, if I cannot reproduce.

I will give it another try.

For the icons seems like the path is broken (likely due to windows \) but this should not have an effect on the label position.

You could do two more things:

1. Check if this also happens, when installing with pip install labelCloud.

2. Send me the label data before and after save (I want to find out, if there is a constant diff).

@ch-sa Thank you for responding.

1. I have installed labelCloud separately with pip and the shift still occurs. This did remove the path error I had before but now I have this error:
   "Could not parse stylesheet of object GUI(0x23842f1b150, name = "MainWindow")"
   Which I think could possibly be due to the same reason (windows \).

2. The label files and related screenshots are posted below for the original imported data, after saving once and after saving twice.
   Original imported data:

After saving once:

label_after_save.txt

After saving twice:

label_after_2_saves.txt

Please let me know if you discover anything.

I fixed the path issue for the icon.

The Qt stylesheets always require a POSIX-like path and no \ even on windows.

For the issue of shifting heights I still could not reproduce it (even on windows).

Though I only tested with other label formats since I did not had calibration files at hand.

Did the issue only appear with the kitti label format or also with other formats?

Ok, now I can replicate it as well.

The issue appears only when using the kitti label format and binary point clouds with calibration files.
So I assume the bug must be somewhere in here:
https://github.com/ch-sa/labelCloud/blob/master/labelCloud/io/labels/kitti.py

Found the issue: Not the correct dimension was used on export with kitti format.
See linked PR: #167

Thanks for reporting.

Fix is merged to master and published to PyPI with v1.1.1.

When I saved the point cloud annotation to the Kitti dataset format, after adding the calibration file, the saving format is still the same as the kitti_untransformed.
like this:Pedestrian 0 0 0 0 0 0 0 1.68 0.7 0.36 -2.2647945 1.10808719 9.39963404 1.57079633
Pedestrian 0 0 0 0 0 0 0 0.15 2.38 0.75 -10.80678875 1.12563597 7.5329695 0.48869219
The position of the 2D box has been 0 and I would like to know how to fix it.
the calibration txt is below(I download in the kitti dataset website, If it works, the position of the 2dbox is not 0，Although the data is not on the Kitti website, it should not be 0.):
P0: 7.215377000000e+02 0.000000000000e+00 6.095593000000e+02 0.000000000000e+00 0.000000000000e+00 7.215377000000e+02 1.728540000000e+02 0.000000000000e+00 0.000000000000e+00 0.000000000000e+00 1.000000000000e+00 0.000000000000e+00
P1: 7.215377000000e+02 0.000000000000e+00 6.095593000000e+02 -3.875744000000e+02 0.000000000000e+00 7.215377000000e+02 1.728540000000e+02 0.000000000000e+00 0.000000000000e+00 0.000000000000e+00 1.000000000000e+00 0.000000000000e+00
P2: 7.215377000000e+02 0.000000000000e+00 6.095593000000e+02 4.485728000000e+01 0.000000000000e+00 7.215377000000e+02 1.728540000000e+02 2.163791000000e-01 0.000000000000e+00 0.000000000000e+00 1.000000000000e+00 2.745884000000e-03
P3: 7.215377000000e+02 0.000000000000e+00 6.095593000000e+02 -3.395242000000e+02 0.000000000000e+00 7.215377000000e+02 1.728540000000e+02 2.199936000000e+00 0.000000000000e+00 0.000000000000e+00 1.000000000000e+00 2.729905000000e-03
R0_rect: 9.999239000000e-01 9.837760000000e-03 -7.445048000000e-03 -9.869795000000e-03 9.999421000000e-01 -4.278459000000e-03 7.402527000000e-03 4.351614000000e-03 9.999631000000e-01
Tr_velo_to_cam: 7.533745000000e-03 -9.999714000000e-01 -6.166020000000e-04 -4.069766000000e-03 1.480249000000e-02 7.280733000000e-04 -9.998902000000e-01 -7.631618000000e-02 9.998621000000e-01 7.523790000000e-03 1.480755000000e-02 -2.717806000000e-01