Spatial image processor
First Claim
1. A spatial image processor neural network for processing an image so as to recognize a whole object from rearranged and partial configurations of component objects, each component object forming a part of the whole object, said network comprising:
- a photo transducer input array for converting a the image to pixel data and sending signals indicative of the pixel data;
a localized gain network module for receiving the signals indicative of the pixel data, increasing the gain of said pixel data in response to attention activations received by said module and outputting localized pixel data;
a parallel memory processor and neuron array for receiving the localized pixel data from said localized gain network module, for processing the localized pixel data into component recognition vectors and peripheral vision object activations;
a component recognition vectors assembly for receiving said component recognition vectors, for receiving associative connections feedback, for generating temporal activations, and for sending feedback data to said parallel memory processor and neuron array;
a temporal spatial retina for receiving the localized pixel data from said localized gain network module and said temporal activations from said component recognition vectors assembly, for generating temporal spatial vectors and for sending signals indicative of said temporal activations;
a temporal parallel memory processor for receiving said temporal spatial vectors from said temporal spatial retina and for generating temporal component recognition vectors; and
a temporal, spatial and object recognition vector array for receiving said temporal component recognition vectors from said temporal parallel memory processor, for forming attention activations to the whole object, for sending prototype object activations of a class of objects to which the whole object belongs, and for sending said associative connections feedback to said component recognition vectors assembly for the component objects recognized as forming a part of the whole object.
1 Assignment
0 Petitions
Accused Products
Abstract
A spatial image processor neural network for processing image data to discriminate between first and second spatial configurations of component objects includes a photo transducer input array for converting an input image to pixel data and sending the data to a localized gain network (LGN) module, a parallel memory processor and neuron array for receiving the pixel data and processing the pixel data into component recognition vectors and chaotic oscillators for receiving the recognition vectors and sending feedback data to the LGN module as attention activations. The network further includes a temporal spatial retina for receiving both the pixel data and temporal feedback activations and generating temporal spatial vectors, which are processed by a temporal parallel processor into temporal component recognition vectors. A spatial recognition vector array receives the temporal component recognition vectors and forms an object representation of the first configuration of component objects.
50 Citations
17 Claims
-
1. A spatial image processor neural network for processing an image so as to recognize a whole object from rearranged and partial configurations of component objects, each component object forming a part of the whole object, said network comprising:
-
a photo transducer input array for converting a the image to pixel data and sending signals indicative of the pixel data;
a localized gain network module for receiving the signals indicative of the pixel data, increasing the gain of said pixel data in response to attention activations received by said module and outputting localized pixel data;
a parallel memory processor and neuron array for receiving the localized pixel data from said localized gain network module, for processing the localized pixel data into component recognition vectors and peripheral vision object activations;
a component recognition vectors assembly for receiving said component recognition vectors, for receiving associative connections feedback, for generating temporal activations, and for sending feedback data to said parallel memory processor and neuron array;
a temporal spatial retina for receiving the localized pixel data from said localized gain network module and said temporal activations from said component recognition vectors assembly, for generating temporal spatial vectors and for sending signals indicative of said temporal activations;
a temporal parallel memory processor for receiving said temporal spatial vectors from said temporal spatial retina and for generating temporal component recognition vectors; and
a temporal, spatial and object recognition vector array for receiving said temporal component recognition vectors from said temporal parallel memory processor, for forming attention activations to the whole object, for sending prototype object activations of a class of objects to which the whole object belongs, and for sending said associative connections feedback to said component recognition vectors assembly for the component objects recognized as forming a part of the whole object. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
-
Specification