Image Segmentation by Complex-Valued Units

Spike synchronisation and de-synchronisation are important for feature binding and separation at various levels in the visual system. We present a model of complex valued neuron activations which are synchronised using lateral couplings. The firing rates of the model neurons correspond to a complex number’s absolute value and obey conventional attractor network relaxation dynamics, while the firing phases correspond to a complex number’s angle and follow the dynamics of a logistic map. During relaxation, we show that features with strong couplings are grouped by firing in the same phase and are separated in phase from features that are coupled weakly or by negative weights. In an example, we apply the model to the level of a hidden representation of an image, thereby segmenting it on an abstract level. We imply that this process can facilitate unsupervised learning of objects in cluttered background.

@InProceedings{WW05, 
 	 author =  {Weber, Cornelius and Wermter, Stefan},  
 	 title = {Image Segmentation by Complex-Valued Units}, 
 	 booktitle = {International Conference on Artificial Neural Networks 2005},
 	 number = {},
 	 volume = {},
 	 pages = {305--310},
 	 year = {2005},
 	 month = {Sep},
 	 publisher = {Springer},
 	 doi = {10.1007/11550822_81}, 
 }