how to understand the result of demo? about posecnn HOT 2 CLOSED

the matrix is [R T] matrix. the R is 33, the T is 31. The unit of values in T is meter.

from posecnn.

I guess I can provide a more intuitive layman's answer:
The matrix is a 3x4 matrix.

• The first 3 columns forms the 3x3 rotation matrix R.
• The last column is the translation T, which represents the distance between the center of the object and the camera in x y z respectively. z is the depth value (always positive), while x and y are all relative to the center of the image (could be negative or positive).

As a simple example, let's take the matrix
[[-0.1556726 -0.98779947 -0.00426031 -0.06283883]
[-0.71000516 0.11489002 -0.69476104 0.09696867]
[ 0.6867741 -0.10513044 -0.7192279 0.7981611 ]]

T is [-0.06283883,0.09696867,0.7981611]. x-value of -0.063 means that the object centroid is 0.063m to the left of the image center, while y-value of 0.097 means that the object centroid is 0.097m below the image center.

from posecnn.