hi, thank you for your outstanding work.
I want to visually check the joint postion on image, I read the joints position from .json file ( I think it's world position in INCH) and read parameters from calibration.json including K(containing f and c) and TW(containing R and T) , I transformed the joints position from world coordinate to pixel coordinate by multipling joints positiong with TW and K in order,
In my understanding, after above transformation, the joints position can be shown directly on image, but when I show the result on image found that the joint position cann't match the image, the joints are far from the actual position on image and the joints are grouped together.
looking into the parameter, I found that the K(containing f and c) is strange compared to normal intrinsic camera parameter, the K(containing f and c) sames like had been normalized, but I don't know how to unnormalized it.
Do you ever try the projection? Thank you very much
hi, thank you for your outstanding work.
I want to visually check the joint postion on image, I read the joints position from .json file ( I think it's world position in INCH) and read parameters from calibration.json including K(containing f and c) and TW(containing R and T) , I transformed the joints position from world coordinate to pixel coordinate by multipling joints positiong with TW and K in order,
In my understanding, after above transformation, the joints position can be shown directly on image, but when I show the result on image found that the joint position cann't match the image, the joints are far from the actual position on image and the joints are grouped together.
looking into the parameter, I found that the K(containing f and c) is strange compared to normal intrinsic camera parameter, the K(containing f and c) sames like had been normalized, but I don't know how to unnormalized it.
Do you ever try the projection? Thank you very much