You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I don't know where else to put this: I think it might be a nice feature to be able to sort out images from cvat or add new ones to a task.
Why: We use CVAT to visualize our data, and pick the training data from it. Sometimes images just dont fit into our dataset. Right now we use a label called "frame-not-used" and mark all unwanted frames with it, then download the task, delete the marked images manually from the dataset and the json and re-upload the task with the fixed json and image dataset.
Being able to add images (maybe even from a different Task with annotations) would be really nice.
Here Again an Overview:
What am I suggesting:
Include data handling tools, like
remove frame from Task
add frame to task
upload frame to task
download all Tasks from a Project as a single JSON/xml/etc
What does this improve?
we can structure our tasks easier and also use CVAT as a persistent data storage for our "living" e.g. training dataset
How can this be implemented?
Add a button to upload/delete single frames/multiple frames to a task
The text was updated successfully, but these errors were encountered:
We are developing a feature to remove images from a task: Deleted frames #4194. Basically you will be able to mark an image as removed. It will be a similar approach which you use, but supported out of the box.
We are not going to add a frame to a task. You can combine tasks to a project. If you need to add data, just create one more task inside the project.
What is the difference between upload and add?
You should be able to export a project as a dataset ("as a single JSON/xml/etc"). What is wrong with the feature?
we can structure our tasks easier and also use CVAT as a persistent data storage for our "living" e.g. training dataset
It is a nice idea. Could you please share information about your company? We are happy to understand better your needs.
We are developing a feature to remove images from a task: Deleted frames #4194. Basically you will be able to mark an image as removed. It will be a similar approach which you use, but supported out of the box.
I wasnt aware that you're working on this feature. This will help me very much!
Dear Community,
I don't know where else to put this: I think it might be a nice feature to be able to sort out images from cvat or add new ones to a task.
Why: We use CVAT to visualize our data, and pick the training data from it. Sometimes images just dont fit into our dataset. Right now we use a label called "frame-not-used" and mark all unwanted frames with it, then download the task, delete the marked images manually from the dataset and the json and re-upload the task with the fixed json and image dataset.
Being able to add images (maybe even from a different Task with annotations) would be really nice.
Here Again an Overview:
What am I suggesting:
The text was updated successfully, but these errors were encountered: