I want to use RGBDSlam on a dataset which is made out of RGB (uint8) and depth (uint16) images from a Kinect. I haven't record the dataset, so I have no access to the rosbags.

I wrote a code that read the images from my disk and publish them on the corresponding topics:

• camera/rgbimage publish a color image CV8U3C
• camera/depthimage publish a grey image CV16U1C
• camera/pointcloud publish a PointCloud2 made out of the pre-aligned RGB and depth image.
  

So far, I've not been able to build a map. Is there any other topic i should publish in order to make it work?

Thanks!