figure denotes the size of the object for object recognition when it is scaled up to the same size as the input image. ... The segmentation operation is implemented by another recurrent network of LTUs, in which each location includes four working units ... It can implement Equations 3.27 and Q by a complex dynamic process of units in the neural network. ... vision Toolkit (iNTV), while a MATLABAr code for the fast image processing method and its LTU network is available in [31 ].
|Title||:||Selective Visual Attention|
|Author||:||Liming Zhang, Weisi Lin|
|Publisher||:||John Wiley & Sons - 2013-03-15|