Contrast enhancement spiking neuron network sensory processing apparatus and methods

• US 9,123,127 B2
• Filed: 12/10/2012
• Issued: 09/01/2015
• Est. Priority Date: 12/10/2012
• Status: Active Grant

• Current Assignee: Brain Corporation
• Original Assignee: Brain Corporation
• Inventors: Richert, Micah
• Primary Examiner(s): Werner, Brian P
• Application Number:
  US13/710,042
• Publication Number:
  US 20140193066A1
• Time in Patent Office: 995 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06N 3/049 Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
• CPC Class Codes: G06T 1/20 Processor architectures; Processor configuration, e.g. pipelining
• CPC Class Codes: G06V 10/451 with interaction between the filter responses, e.g. cortical complex cells
• CPC Class Codes: G06V 10/955 using specific electronic processors

Abstract

Apparatus and methods for contrast enhancement and feature identification. In one implementation, an image processing apparatus utilizes latency coding and a spiking neuron network to encode image brightness into spike latency. The spike latency is compared to a saliency window in order to detect early responding neurons. Salient features of the image are associated with the early responding neurons. A inhibitory neuron receives salient feature indication and provides inhibitory signal to the other neurons within an area of influence of the inhibitory neuron. The inhibition signal reduces probability of responses by the other neurons to stimulus that is proximate to the feature thereby increasing contrast within the encoded data. The contrast enhancement may facilitate feature identification within the image. Feature detection may be used for example for image compression, background removal and content distribution.

Claims

1. A computerized apparatus comprising a storage medium, the storage medium comprising a plurality of computer-readable instructions configured to, when executed, detect an object in a sensory input by at least:

based on at least one feed-forward signal, generate an inhibitory signal configured to suppress generation of a secondary feed-forward signal, thereby enabling the object detection;

wherein;

the at least one feed-forward signal comprises one or more pulses generated based on a portion of the sensory input comprising a representation of the object;

the secondary feed-forward signal corresponds to a portion of the sensory input proximate the representation of the object; and

the representation of the object being characterized by a parameter having a value that is different from a parameter value associated with the portion of the sensory input.

2. The apparatus of claim 1, wherein:

the sensory input comprises an image comprising the representation of the object;

the feed-forward signal comprises one or more pulses having a first range of latencies associated therewith; and

the secondary feed-forward signal is configured based on a second range of latencies, the latencies within the second range being greater than the latencies within the first range.

3. The apparatus of claim 2, wherein:

the parameter is configured to characterize a luminance of pixels within the image; and

individual latencies within the first range of latencies are encoded based on a logarithmic function of the luminance of pixels.

4. The apparatus of claim 2, wherein an inhibition indication is effectuated at least by an inhibition trace, the inhibition trace being incremented responsive to an occurrence of a pulse latency of a first plurality of pulses, a second plurality of pulses being within a latency window.

5. The apparatus of claim 2, wherein:

the one or more pulses comprise a first and a second pulse associated with a first and a second pixel, respectively, each characterized by a pixel luminance being within an object luminance range; and

a pixel luminance of the portion of the image that is outside the object luminance range.

6. The apparatus of claim 2, wherein:

the inhibitory signal is generated based on another feed-forward signal being generated based on another image comprising another representation of the object, the another image preceding the image; and

a time interval between the image and the another image is greater than the first range of latencies.

7. The apparatus of claim 1, wherein:

the sensory input comprises an image comprising the representation of the object;

the inhibitory signal is configured to transform the image into a spiking representation comprising the at least one feed-forward signal;

the inhibitory signal is characterized by an inhibition area within the image;

the representation of the object is characterized by an object area; and

the spiking representation is characterized by absence of spikes corresponding to pixels of the image that are outside the object area and within the inhibition area.

8. The apparatus of claim 1, wherein:

the sensory input comprises two or more frames characterized by an inter-frame interval, individual ones of the two or more frames comprising the representation of the object and an object surrounding area; and

the suppression of the generation of the secondary feed-forward signal is characterized by an absence of spikes associated with one or more pixels within the object surrounding area during at least two inter-frame intervals.

9. The apparatus of claim 1, wherein:

the sensory input comprises two or more frames characterized by an inter-frame interval, individual ones of the two or more frames comprising the representation of the object and an object surrounding area; and

the suppression of the generation of the secondary feed-forward signal is configured to reduce a number of spikes associated with one or more pixels within the object surrounding area during at least two inter-frame intervals.

10. The apparatus of claim 1, wherein:

the sensory input comprises first and second frames characterized by an inter-frame interval, individual ones of the first and the second frames comprising a first and a second representation of the object, respectively, and a representation of a surrounding area;

the at least one feed-forward signal comprises a first feed-forward signal and a second feed-forward signal generated based on the first and the second representation of the object, respectively;

the inhibitory signal is configured based on the first and the second feed-forward signal.

11. A computerized method of increasing a contrast of an image comprising a plurality of pixels using at least a spiking neuron network, the method comprising:

based on a feed-forward stimulus representative of an object within the image, generating a spike output by one or more excitatory neurons of the network; and

based on the spike output, generating an inhibitory signal by an inhibitory neuron;

wherein;

individual ones of the plurality of pixels are characterized by a pixel parameter;

the object is characterized by at least;

(i) an object spatial extent having an object pixel parameter associated therewith, and (ii) a surround spatial extent having a surround pixel parameter associated therewith, the surround pixel parameter being different from the object pixel parameter; and

the inhibitory signal is configured to suppress a surrounding spike output associated with a plurality of pixels within the surround spatial extent, thereby effectuating the contrast increase.

12. The method of claim 11, wherein the pixel parameter comprises a pixel brightness range, and the inhibitory signal is configured to increase a difference between a network response to pixels of the plurality of pixels within the object spatial extent and the plurality of pixels within the surround spatial extent.

13. The method of claim 11, wherein:

the pixel parameter comprises a pixel color;

the object pixel parameter corresponds to a first color; and

the surround pixel parameter corresponds to a second color being different from the first color.

14. The method of claim 11, wherein:

the inhibitory neuron is characterized by an area of influence corresponding to an area of the image, the area of influence spatially overlapping the surrounding area; and

the inhibitory signal is configured to suppress the spike output based on pixels of the plurality of pixels within the area of influence.

15. The method of claim 11, further comprising, based on another feed-forward stimulus representative of another object within the image, generating a spike output by another one or more excitatory neurons of the network;

wherein

the another object is characterized by another object spatial extent having another object pixel parameter associated therewith;

the another object spatial extent is configured to be non-overlapping with the surround spatial extent; and

the another object pixel parameter and the surround pixel parameter fall in a parameter value range from the object pixel parameter.

16. The method of claim 15, wherein a difference between the another object pixel parameter and the surround pixel parameter is smaller in value as compared to a difference between the another object pixel parameter and the object pixel parameter.

17. The method of claim 11, wherein the suppression of the surround spike output is configured based at least in part on a number of spikes in the spike output.

18. A spiking neuron network sensory processing system, comprising:

a plurality of excitatory neurons configured to encode a sensory stimulus into a plurality of pulses; and

at least one inhibitory neuron configured to provide an inhibitory indication to at least a portion of the plurality of excitatory neurons via one or more inhibitory connections; and

wherein;

the inhibitory indication is based at least in part on two or more of the plurality of pulses being received by the at least one inhibitory neuron over one or more feed-forward connections;

the inhibitory indication is configured to prevent at least one neuron within the portion from generating a response during a stimulus interval subsequent to the provision of the inhibitory indication;

the sensory stimulus comprises two or more frames of pixels, at least one of the two or more frames comprising a representation of an object;

individual ones of the two or more frames of pixels comprised of pixels characterized by a parameter; and

the inhibitory indication is generated based on a pixel parameter within the representation of the object being different from a pixel parameter outside the object representation.

19. The system of claim 18, wherein:

the inhibitory indication is characterized by an inhibition trace, the inhibition trace configured to be;

incremented responsive to the receipt of the two or more of the plurality of pulses;

decayed with time subsequent to the increment; and

wherein the prevention of the generation of the response is based at least in part on a value of the inhibition trace exceeding an inhibitory threshold.

20. The system of claim 18, wherein:

a feed-forward sensory stimulus comprises a background portion comprising an uncorrelated sensory input component;

the feed-forward sensory stimulus further comprises a coherent signal component;

the plurality of pulses constitute an encoded input; and

the inhibitory indication is capable of preventing the uncorrelated sensory input component from being encoded into the plurality of pulses, thereby effectuating an increased signal-to-noise ratio of the encoded input.