


	An Efficient Direct Method for Improving visual SLAM
		  G. Silveira, E. Malis, P. Rives


-----------------------------------------------------------------------
			   SUMMARY.TXT
-----------------------------------------------------------------------


This extended object is composed by two visual SLAM results, which are
discussed in the paper.

In both demos, the left frame shows the input images superposed with
the tracked regions, while the right frame shows both the pose and
scene being incrementally reconstructed.

As expressed in the title, the results are obtained by a direct method
(no image features are used). This means that all computations are
performed using pixel intensities directly. Also to respect real-time
constraints, a suitable selection of which image regions are
considered for processing is carried out.

