Image processing system for extraction of contextual information and associated methods

Patent metadata

Field	Value
Publication number	US-9830527-B2
Application number	US-201514593413-A
Country	US
Kind code	B2
Filing date	Jan 9, 2015
Priority date	Jan 9, 2015
Publication date	Nov 28, 2017
Grant date	Nov 28, 2017

Abstract

An image processing system includes a first processor that acquires frames of image data. For each frame of data, the first processor generates a Gaussian pyramid for the frame of data, extract histogram of oriented gradient (HOG) descriptors for each level of the Gaussian pyramid, compresses the HOG descriptors, and sends the compressed HOG descriptors. A second processor is coupled to the first processor and is configured to receive the compressed HOG descriptors, aggregate the compressed HOG descriptors into windows, compare data of each window to at least one stored model, and generate output based upon the comparison.

First claim

The invention claimed is:

1. A system situated in a vehicle, comprising:
a first processor situated in the vehicle and configured to:
acquire frames of image data, and
for each frame of data,
generate a Gaussian pyramid for the frame of data,
extract histogram of oriented gradient (HOG) descriptors for each level of the Gaussian pyramid, and separate the HOG descriptor for each level into a plurality of bins comprising contrast sensitive bins, contrast insensitive bins, and gradient energy bins,
compress the HOG descriptors by selecting each contrast sensitive bin and each contrast insensitive bin, binarizing each HOG descriptor for each level such that each selected bin holds a binary value, and generating the compressed HOG descriptors as the binary values of the selected bins, wherein binarizing each HOG descriptor is performed by determining the binary values for the plurality of bins corresponding to that HOG descriptor by comparing each element of that HOG descriptor to a threshold value and generating the corresponding binary value based upon the comparison, and
send the compressed HOG descriptors;

a second processor situated in the vehicle, coupled to the first processor, and configured to:
receive the compressed HOG descriptors,
aggregate the compressed HOG descriptors into windows,
compare data of each window to at least one stored model, and
generate output based upon the comparison.

2. The system of claim 1 , wherein the output indicates an impending physical collision.

3. The system of claim 1 , wherein the first processor is configured to select some of the plurality of bins by selecting each of the plurality of bins.

4. A system situated in a vehicle, comprising:
a first processor situated in the vehicle and configured to:
acquire frames of image data, and
for each frame of data,
generate a Gaussian pyramid for the frame of data,
extract histogram of oriented gradient (HOG) descriptors for each level of the Gaussian pyramid, and separate the HOG descriptor for each level into a plurality of bins comprising contrast sensitive bins, contrast insensitive bins, and gradient energy bins,
compress the HOG descriptors by selecting each contrast sensitive bin and each gradient energy bin, binarizing each HOG descriptor for each level such that each selected bin holds a binary value, and generating the compressed HOG descriptors as the binary values of the selected bins, wherein binarizing each HOG descriptor is performed by determining the binary values for the plurality of bins corresponding to that HOG descriptor by comparing each element of that HOG descriptor to a threshold value and generating the corresponding binary value based upon the comparison, and
send the compressed HOG descriptors;

a second processor situated in the vehicle, coupled to the first processor, and configured to:
receive the compressed HOG descriptors,
aggregate the compressed HOG descriptors into windows,
compare data of each window to at least one stored model, and
generate output based upon the comparison.

5. A system situated in a vehicle, comprising:
a first processor situated in the vehicle and configured to:
acquire frames of image data, and
for each frame of data,
generate a Gaussian pyramid for the frame of data,
extract histogram of oriented gradient (HOG) descriptors for each level of the Gaussian pyramid, and separate the HOG descriptor for each level into a plurality of bins comprising contrast sensitive bins, contrast insensitive bins, and gradient energy bins,
compress the HOG descriptors by selecting each contrast sensitive bin, binarizing each HOG descriptor for each level such that each selected bin holds a binary value, and generating the compressed HOG descriptors as the binary values of the selected bins, wherein binarizing each HOG descriptor is performed by determining the binary values for the plurality of bins corresponding to that HOG descriptor by comparing each element of that HOG descriptor to a threshold value and generating the corresponding binary value based upon the comparison, and
send the compressed HOG descriptors;

a second processor situated in the vehicle, coupled to the first processor, and configured to:
receive the compressed HOG descriptors,
aggregate the compressed HOG descriptors into windows,
compare data of each window to at least one stored model, and
generate output based upon the comparison.

6. A system situated in a vehicle, comprising:
a first processor situated in the vehicle and configured to:
acquire frames of image data, and
for each frame of data,
generate a Gaussian pyramid for the frame of data,
extract histogram of oriented gradient (HOG) descriptors for each level of the Gaussian pyramid, and separate the HOG descriptor for each level into a plurality of bins comprising contrast sensitive bins, contrast insensitive bins, and gradient energy bins,
compress the HOG descriptors by selecting each contrast insensitive bin and each gradient energy bin, binarizing each HOG descriptor for each level such that each selected bin holds a binary value, and generating the compressed HOG descriptors as the binary values of the selected bins, wherein binarizing each HOG descriptor is performed by determining the binary values for the plurality of bins corresponding to that HOG descriptor by comparing each element of that HOG descriptor to a threshold value and generating the corresponding binary value based upon the comparison, and
send the compressed HOG descriptors;

a second processor situated in the vehicle, coupled to the first processor, and configured to:
receive the compressed HOG descriptors,
aggregate the compressed HOG descriptors into windows,
compare data of each window to at least one stored model, and
generate output based upon the comparison.

7. A system situated in a vehicle, comprising:
a first processor situated in the vehicle and configured to:
acquire frames of image data, and
for each frame of data,
generate a Gaussian pyramid for the frame of data,
extract histogram of oriented gradient (HOG) descriptors for each level of the Gaussian pyramid, and separate the HOG descriptor for each level into a plurality of bins comprising contrast sensitive bins, contrast insensitive bins, and gradient energy bins,
compress the HOG descriptors by selecting each contrast insensitive bin, binarizing each HOG descriptor for each level such that each selected bin holds a binary value, and generating the compressed HOG descriptors as the binary values of the selected bins, wherein binarizing each HOG descriptor is performed by determining the binary values for the plurality of bins corresponding to that HOG descriptor by comparing each element of that HOG descriptor to a threshold value and generating the corresponding binary value based upon the comparison, and
send the compressed HOG descriptors;

a second processor situated in the vehicle, coupled to the first processor, and configured to:
receive the compressed HOG descriptors,
aggregate the compressed HOG descriptors into windows,
compare data of each window to at least one stored model, and
generate output based upon the comparison.

8. A system situated in a vehicle, comprising:
a first processor situated in the vehicle and configured to:
acquire frames of image data, and
for each frame of data,
generate a Gaussian pyramid for the frame of data,
extract histogram of oriented gradient (HOG) descriptors for each level of the Gaussian pyramid, and separate the HOG descriptor for each level into a plurality of bins comprising contrast sensitive bins, contrast insensitive bins, and gradient energy bins, wherein each bin corresponds to an angle over which each HOG descrip

Assignees

• St Microelectronics Srl
• Univ Degli Studi Di Milano—Bicocca

Inventors

• Margari Alberto
• Pau Danilo Pietro
• Schettini Raimondo

Classifications

• G06V20/58 Primary

  Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads · CPC title

• G06V10/50

  by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis · CPC title

• G06K9/4642 Primary

  Physics · mapped topic

• G06K9/00973

  Physics · mapped topic

• G06K9/00805

  Physics · mapped topic