Unsupervised Learning of Overlapping Image Components Using Divisive Input Modulation
Figure 4
Errors in parsing the overlapping squares
tasks with (a) , (b) , and (c) ; , and in each case.
Each bar shows the proportion of errors generated across 1000 test images. Each
bar is subdivided into the proportion of false negatives (lighter, lower,
section) and the proportion of false positives (darker, upper, section).