The results below show the confusion matrices for decoding the identity of objects from the isolated object trials. In general, there is not a clear pattern of mistakes made by the classifier, even on the 16 object stimulus set where the images come for 4 categories. There are however, a few trends in the 16 object dataset, such as the fact that the image of an apple was often mistaken as an image of an orange, there is more confusion among the couches, and the cars generally had lower decoding accuracies than the other categories of objects.
