I visualized camera poses and positions provided by this repo using viser.
As shown in this visualization result, cameras are facing in the opposite direction of the region of interest.
In addition, unlike Wildtrack and GMVD datasets, the Z-axis representing the height in the world coordinate system is also inverted.
(Camera positions in the world coordinate system appear to be correct.)
Are the camera extrinsic parameters correct?

Visualization of MultiviewX

Visualization of Wildtrack

Visualization of GMVD