3D point clouds and 2D images cannot be annotated together

cvat-ai / cvat

Annotate better with CVAT, the industry-leading data engine for machine learning. Used and trusted by teams at any scale, for data of any scale.

https://cvat.ai

MIT License

12.7k stars 3.02k forks source link

3D point clouds and 2D images cannot be annotated together #6417

Open quanhuanhuan opened 1 year ago

quanhuanhuan commented 1 year ago

After the 3D point cloud is marked, the wire-frame data corresponding to the 2D image is not automatically displayed https://opencv.github.io/cvat/docs/manual/advanced/contextual-images/#folder-structure

bsekachev commented 1 year ago

Hi @quanhuanhuan

You are right. Unfortunately right now there is not relation between 3D scene and 2D related images. Camera calibration info is not used.

bsekachev commented 1 year ago

Related #4810

quanhuanhuan commented 1 year ago

hi @bsekachev, thank you for your answer

https://opencv.github.io/cvat/docs/manual/advanced/contextual-images/#contextual-images From the current point of view, this function provided by cvat is not useful, then does cvat plan to support the function of 3D and 2D joint annotation in the future?

Kuroneko089 commented 1 year ago

It would be great to implement this function, it is super useful.

By the way, in 3D scene, the orientation of an object should also be annotated in some way. At present, two objects with opposite orientations could get the same yaw result. I suggest to draw a cross at the front side of an object.

zhoupan163 commented 1 year ago

It would be great to implement this function, it is super useful.

By the way, in 3D scene, the orientation of an object should also be annotated in some way. At present, two objects with opposite orientations could get the same yaw result. I suggest to draw a cross at the front side of an object.

I can't agree more