4 thoughts on “Horse Head Reconstruction

    1. Christoph Heindl Post author

      Hi and thanks for clarifying that. Just found out the details at


      What do you mean with “Second advantage is that ReconstructMe has access to the accelerometer inside the kinect, so the number of rotational degrees of freedom is reduced to one.” ?

      We do not use any additional sensor except the visual data from the sensor. Otherwise it wouldn’t work with different sensors like the ASUS Xtion.

      1. unicorniversity

        Ohh, didn’t know that. Just assumed it would, from our own experiments with kinect raw data, where we relied heavily on the accelerometer’s data, smoothed by a kalman filter. In simple terms we thought as this sensor was there, you’d use it.
        In retroperspective that explains why different scans of the same are often quite missaligned in all three axis.
        That means the definition of the scanning-volume refers to the initial position of the camera?

        1. Christoph Heindl Post author

          Yes, the first time you press ‘p’ the volume is fixed. Use the multiscan feature to align scans over multiple volumes.


Leave a Reply