Histograms of local features for the object dataset

The acquisition was performed and is described in more details by Lyubova and Filliat. Each image contains one visual object and possibly the hand of the operator.

Images contains objects from a subset of 10 objects each appearing in more than a thousand frames.During the acquisition, the objects are moved and rotated. Hence they are presented from distinct points of view and they may be partially cluttered by the hand of the operator. For more information on the format please look at github.com/omangin/multimodal/blob/master/db/models/objects.py.