Replication of directionally selective pathways in the retina using simulated artificial neural networks

The development of an artificial directionally selective pathway based on neural network architecture provides a critical element to the development of a complete artificial vision system. By developing algorithms which emulate the neural processing pathways within the retina, retinal implants can be developed which not only encode light in a feedforward manner but also include feedback mechanisms and computational pathways. In addition, the application of neural networks to the replication of neural pathways provides a more natural solution than by replicating these pathways in the conventional serial manner. Parallelism is the key to neurological pathway replication because biological neural networks are naturally parallel. Ultimately neural data flow computers will likely provide the best solution to biological replication do to the vast similarities between the electrical hardware and the biological architecture.