Annotate in 2D RGB image is easy since we can use many toolkit to do this task such as "Labelme", but when it comes to 3D lane det, we have to annotate it in 2D RGB image first (including N points for example), than calculate the distance between the 3D projected points and the 2D lanes and choose the points whose distance are less than the certain threshold, does these means that the 3D annotation will have N points as well?
And could you please provide an annotation tool for 3D lane det so that I can annotate my own dataset?
Annotate in 2D RGB image is easy since we can use many toolkit to do this task such as "Labelme", but when it comes to 3D lane det, we have to annotate it in 2D RGB image first (including N points for example), than calculate the distance between the 3D projected points and the 2D lanes and choose the points whose distance are less than the certain threshold, does these means that the 3D annotation will have N points as well?
And could you please provide an annotation tool for 3D lane det so that I can annotate my own dataset?