High-performance sensor fusion architecture
First Claim
1. A computer implemented method of object detection comprising an act of causing a processor to perform operations of:
- receiving images of an area occupied by at least one object;
extracting image features including wavelet features from the images;
classifying the image features to produce object class confidence data, wherein the classifying operation is performed by at least two sub-classifiers; and
performing data fusion on the object class confidence data to produce a detected object estimate, wherein the operation of performing data fusion comprises the operations of;
initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers;
training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and
using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown;
wherein the operation of classifying image features comprises processing the image features with at least one classification algorithm; and
wherein at least one of the classification algorithms is selected from the group consisting of a trained C5 decision tree, a trained Nonlinear Discriminant Analysis network, and a trained Fuzzy Aggregation Network.
1 Assignment
0 Petitions
Accused Products
Abstract
A vision-based system for automatically detecting the type of object within a specified area, such as the type of occupant within a vehicle is presented. The type of occupant can then be used to determine whether an airbag deployment system should be enabled or not. The system extracts different features, including wavelet features and/or a disparity map from images captured by image sensors. These features are then processed by classification algorithms to produce class confidences for various occupant types. The occupant class confidences are fused and processed to determine occupant type. In a preferred embodiment, image features from image edges, wavelet features, and disparity are used. Various classification algorithms may be implemented to classify the object. Use of the disparity map and/or wavelet features provides greater computational efficiency.
79 Citations
144 Claims
-
1. A computer implemented method of object detection comprising an act of causing a processor to perform operations of:
-
receiving images of an area occupied by at least one object; extracting image features including wavelet features from the images; classifying the image features to produce object class confidence data, wherein the classifying operation is performed by at least two sub-classifiers; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the operation of performing data fusion comprises the operations of; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown; wherein the operation of classifying image features comprises processing the image features with at least one classification algorithm; and wherein at least one of the classification algorithms is selected from the group consisting of a trained C5 decision tree, a trained Nonlinear Discriminant Analysis network, and a trained Fuzzy Aggregation Network.
-
-
2. A computer implemented method of object detection comprising an act of causing a processor to perform operations of:
-
receiving images of an area occupied by at least one object; extracting image features including wavelet features from the images; classifying the image features to produce object class confidence data, wherein the classifying operation is performed by at least two sub-classifiers; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the operation of performing data fusion comprises the operations of; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown; wherein the operation of classifying image features comprises processing the image features with at least one classification algorithm; wherein the operation of extracting image features comprises the operation of; extracting wavelet coefficients of the at least one object occupying an area of the images; and
wherein the operation of classifying the image features comprises processing the wavelet coefficients with at least one classification algorithm to produce object class confidence data; andwherein at least one of the classification algorithms is selected from the group consisting of a trained C5 decision tree, a trained Nonlinear Discriminant Analysis network, and a trained Fuzzy Aggregation Network. - View Dependent Claims (3, 4, 5)
-
-
6. A computer implemented method of object detection comprising an act of causing a processor to perform operations of:
-
receiving images of an area occupied by at least one object; extracting image features including wavelet features from the images; classifying the image features to produce object class confidence data, wherein the classifying operation is performed by at least two sub-classifiers; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the operation of performing data fusion comprises the operations of; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown; wherein the operation of classifying image features comprises processing the image features with at least one classification algorithm; wherein the operation of extracting image features comprises the operation of; extracting wavelet coefficients of the at least one object occupying an area of the images; and
wherein the operation of classifying the image features comprises processing the wavelet coefficients with at least one classification algorithm to produce object class confidence data; andwherein the operation of extracting image features further comprises the operations of; detecting edges of the at least one object within the images; masking the edges with a background mask to find important edges; calculating edge pixels from the important edges; and producing edge density maps from the important edges, the edge density map providing the image features, and wherein the operation of classifying the image features comprises processing the edge density map with at least one of the classification algorithms to produce object class confidence data. - View Dependent Claims (7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
-
-
18. A computer implemented method of object detection comprising an act of causing a processor to perform operations of:
-
receiving images of an area occupied by at least one object; extracting image features including wavelet features from the images; classifying the image features to produce object class confidence data, wherein the classifying operation is performed by at least two sub-classifiers; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the operation of performing data fusion comprises the operations of; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as Outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown; wherein the operation of classifying image features comprises processing the image features with at least one classification algorithm; wherein the operation of extracting image features comprises the operation of; extracting wavelet coefficients of the at least one object occupying an area of the images; and
wherein the operation of classifying the image features comprises processing the wavelet coefficients with at least one classification algorithm to produce object class confidence data; andwherein the operation of extracting image features further comprises the operations of; receiving a stereoscopic pair of images of an area occupied by at least one object; detecting pattern regions and non-pattern regions within each of the pair of images using a texture filter; generating an initial estimate of spatial disparities between the pattern regions within each of the pair of images; using the initial estimate to generate a subsequent estimate of the spatial disparities between the non-pattern regions based on the spatial disparities between the pattern regions using disparity constraints; and
wherein the disparity constraints comprise consistency requirement constraints and local smoothness constraints, wherein the consistency requirement constraints are defined by - View Dependent Claims (19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29)
-
-
30. A computer implemented method of object detection comprising an act of causing a processor to perform operations of:
-
receiving images of an area occupied by at least one object; extracting image features including wavelet features from the images; classifying the image features to produce object class confidence data, wherein the classifying operation is performed by at least two sub-classifiers; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the operation of performing data fusion comprises the operations of; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown where the at least one object comprises a vehicle occupant and the area comprises a vehicle occupancy area, and further comprising an operation of processing the detected object estimate to provide signals to vehicle systems; wherein the signals comprise airbag enable and disable signals; wherein the method further comprises an operation of capturing images from a sensor selected from a group consisting of CMOS vision sensors and CCD vision sensors; and wherein at least one of the classification algorithms is selected from the group consisting of a trained C5 decision tree, a trained Nonlinear Discriminant Analysis network, and a trained Fuzzy Aggregation Network.
-
-
31. A computer implemented method of object detection comprising an act of causing a processor to perform operations of:
-
receiving a stereoscopic pair of images of an area occupied by at least one object; extracting image features from the images, with at least a portion of the image features being extracted by the operations of; detecting pattern regions and non-pattern regions within each of the pair of images using a texture filter; generating an initial estimate of spatial disparities between the pattern regions within each of the pair of images; using the initial estimate to generate a subsequent estimate of the spatial disparities between the non-pattern regions based on the spatial disparities between the pattern regions using disparity constraints; iteratively using the subsequent estimate as the initial estimate in the operation of using the initial estimate to generate a subsequent estimate in order to generate further subsequent estimates of the spatial disparities between the non-pattern regions based on the spatial disparities between the pattern regions using the disparity constraints until there is no change between the results of subsequent iterations, thereby generating a final estimate of the spatial disparities; generating a disparity map of the area occupied by at least one object from the final estimate of the spatial disparities; classifying the image features to produce object class confidence data, wherein the classifying operation is performed by at least two sub-classifiers and with at least a portion of the classifying being performed by processing the disparity map with at least one classification algorithm to produce object class confidence data; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the operation of performing data fusion comprises the operations of; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown. - View Dependent Claims (32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48)
-
-
49. A computer program product for object detection, the computer program product comprising means, stored on a computer readable medium, for:
-
receiving images of an area occupied by at least one object; extracting image features including wavelet features from the images; classifying the image features to produce object class confidence data, wherein the classifying means use at least two sub-classifiers; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the means for performing data fusion comprises the means for; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown; wherein the means for classifying image features comprises a means for processing the image features with at least one classification algorithm; and wherein at least one of the classification algorithms is selected from the group consisting of a trained C5 decision tree, a trained Nonlinear Discriminant Analysis network, and a trained Fuzzy Aggregation Network.
-
-
50. A computer program product for object detection, the computer program product comprising means, stored on a computer readable medium, for:
-
receiving images of an area occupied by at least one object; extracting image features including wavelet features from the images; classifying the image features to produce object class confidence data, wherein the classifying means use at least two sub-classifiers; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the means for performing data fusion comprises the means for; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown; wherein the means for extracting image features comprises a means for; extracting wavelet coefficients of the at least one object in the images; and
wherein the means for classifying the image features comprises processing the wavelet coefficients with at least one classification algorithm to produce object class confidence data; andwherein at least one of the classification algorithms is selected from the group consisting of a trained C5 decision tree, a trained Nonlinear Discriminant Analysis network, and a trained Fuzzy Aggregation Network. - View Dependent Claims (51, 52, 53)
-
-
54. A computer program product for object detection, the computer program product comprising means, stored on a computer readable medium, for:
-
receiving images of an area occupied by at least one object; extracting image features including wavelet features from the images; classifying the image features to produce object class confidence data, wherein the classifying means use at least two sub-classifiers; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the means for performing data fusion comprises the means for; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown; wherein the means for extracting image features comprises a means for; extracting wavelet coefficients of the at least one object in the images; and
wherein the means for classifying the image features comprises processing the wavelet coefficients with at least one classification algorithm to produce object class confidence data; andwherein the means for extracting image features further comprises means for; detecting edges of the at least one object within the images; masking the edges with a background mask to find important edges; calculating edge pixels from the important edges; and producing edge density maps from the important edges, the edge density map providing the image features, and wherein the means for classifying the image features processes the edge density map with at least one of the classification algorithms to produce object class confidence data. - View Dependent Claims (55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65)
-
-
56. A computer program product for object detection as set forth in claim 55, further comprising means for:
-
detecting motion of the at least one object within the images; calculating motion pixels from the motion; and producing motion density maps from the motion pixels, the motion density map providing the image features; and wherein the means for classifying the image features processes the motion density map with at least one of the classification algorithms to produce object class confidence data, which is used independently of the data fusion to produce an independent detected object estimate.
-
-
57. A computer program product for object detection as set forth in claim 56, where the at least one object comprises a vehicle occupant and the area comprises a vehicle occupancy area, and further comprising a means for processing the detected object estimate to provide signals to vehicle systems.
-
58. A computer program product for object detection as set forth in claim 57, wherein the signals comprise airbag enable and disable signals.
-
59. A computer program product for object detection as set forth in claim 58, wherein the computer program product further comprises a means for capturing images from a sensor selected from a group consisting of CMOS vision sensors and CCD vision sensors.
-
60. A computer program product for object detection as set forth in claim 59, wherein at least one of the classification algorithms is selected from the group consisting of a trained C5 decision tree, a trained Nonlinear Discriminant Analysis network, and a trained Fuzzy Aggregation Network.
-
61. A computer program product for object detection as set forth in claim 54, further comprising means for:
-
detecting motion of the at least one object within the images; calculating motion pixels from the motion; and producing motion density maps from the motion pixels, the motion density map providing the image features; and wherein the means for classifying the image features processes the motion density map with at least one of the classification algorithms to produce object class confidence data.
-
-
62. A computer program product for object detection as set forth in claim 61, where the at least one object comprises a vehicle occupant and the area comprises a vehicle occupancy area, and further comprising a means for processing the detected object estimate to provide signals to vehicle systems.
-
63. A computer program product for object detection as set forth in claim 62, wherein the signals comprise airbag enable and disable signals.
-
64. The computer program product of claim 63, wherein the computer program product further comprises a means for capturing images from a sensor selected from a group consisting of CMOS vision sensors and CCD vision sensors.
-
65. A computer program product for object detection as set forth in claim 64, wherein at least one of the classification algorithms is selected from the group consisting of a trained C5 decision tree, a trained Nonlinear Discriminant Analysis network, and a trained Fuzzy Aggregation Network.
-
66. A computer program product for object detection, the computer program product comprising means, stored on a computer readable medium, for:
-
receiving images of an area occupied by at least one object; extracting image features including wavelet features from the images; classifying the image features to produce object class confidence data, wherein the classifying means use at least two sub-classifiers; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the means for performing data fusion comprises the means for; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown; wherein the means for extracting image features comprises a means for; extracting wavelet coefficients of the at least one object in the images; and
wherein the means for classifying the image features comprises processing the wavelet coefficients with at least one classification algorithm to produce object class confidence data; andwherein the means for extracting image features further comprises means for; receiving a stereoscopic pair of images of an area occupied by at least one object; detecting pattern regions and non-pattern regions within each of the pair of images using a texture filter; generating an initial estimate of spatial disparities between the pattern regions within each of the pair of images; using the initial estimate to generate a subsequent estimate of the spatial disparities between the non-pattern regions based on the spatial disparities between the pattern regions using disparity constraints; and
wherein the disparity constraints comprise consistency requirement constraints and local smoothness constraints, wherein the consistency requirement constraints are defined by - View Dependent Claims (67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77)
-
-
78. A computer program product for object detection, the computer program product comprising means, stored on a computer readable medium, for:
-
receiving images of an area occupied by at least one object; extracting image features including wavelet features from the images; classifying the image features to produce object class confidence data, wherein the classifying means use at least two sub-classifiers; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the means for performing data fusion comprises the means for; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown; where the at least one object comprises a vehicle occupant and the area comprises a vehicle occupancy area, and further comprising a means for processing the detected object estimate to provide signals to vehicle systems; wherein the signals comprise airbag enable and disable signals; wherein the computer program product further comprises a means for capturing images from a sensor selected from a group consisting of CMOS vision sensors and CCD vision sensors; and wherein at least one of the classification algorithms is selected from the group consisting of a trained C5 decision tree, a trained Nonlinear Discriminant Analysis network, and a trained Fuzzy Aggregation Network.
-
-
79. A computer program product for object detection, the computer program product comprising means, stored on a computer readable medium, for:
-
receiving a stereoscopic pair of images of an area occupied by at least one object; extracting image features from the images, with at least a portion of the image features being extracted by means for; detecting pattern regions and non-pattern regions within each of the pair of images using a texture filter; generating an initial estimate of spatial disparities between the pattern regions within each of the pair of images; using the initial estimate to generate a subsequent estimate of the spatial disparities between the non-pattern regions based on the spatial disparities between the pattern regions using disparity constraints; iteratively using the subsequent estimate as the initial estimate in the means for using the initial estimate to generate a subsequent estimate in order to generate further subsequent estimates of the spatial disparities between the non-pattern regions based on the spatial disparities between the pattern regions using the disparity constraints until there is no change between the results of subsequent iterations, thereby generating a final estimate of the spatial disparities; generating a disparity map of the area occupied by at least one object from the final estimate of the spatial disparities; classifying the image features to produce object class confidence data, wherein the classifying means use at least two sub-classifiers and with at least a portion of the classifying being performed by processing the disparity map with at least one classification algorithm to produce object class confidence data; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the means for performing data fusion comprises the means for; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown. - View Dependent Claims (80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96)
-
-
97. An apparatus for object detection comprising a computer system including a processor, a memory coupled with the processor, an input coupled with the processor for receiving images, and an output coupled with the processor for outputting information based on an object estimation, wherein the computer system further comprises means, residing in its processor and memory, for:
-
receiving images of an area occupied by at least one object; extracting image features including wavelet features from the images; classifying the image features to produce object class confidence data, wherein the classifying means use at least two sub-classifiers; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the means for performing data fusion comprises the means for; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown; wherein the means for classifying image features comprises a means for processing the image features with at least one classification algorithm; and wherein at least one of the classification algorithms is selected from the group consisting of a trained C5 decision tree, a trained Nonlinear Discriminant Analysis network, and a trained Fuzzy Aggregation Network.
-
-
98. An apparatus for object detection comprising a computer system including a processor, a memory coupled with the processor, an input coupled with the processor for receiving images, and an output coupled with the processor for outputting information based on an object estimation, wherein the computer system further comprises means, residing in its processor and memory, for:
-
receiving images of an area occupied by at least one object; extracting image features including wavelet features from the images; classifying the image features to produce object class confidence data, wherein the classifying means use at least two sub-classifiers; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the means for performing data fusion comprises the means for; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown; wherein means for extracting image features comprises a means for; extracting wavelet coefficients of the at least one object in the images; and wherein the means for classifying the image features comprises processing the wavelet coefficients with at least one classification algorithm to produce object class confidence data; and wherein at least one of the classification algorithms is selected from the group consisting of a trained C5 decision tree, a trained Nonlinear Discriminant Analysis network, and a trained Fuzzy Aggregation Network. - View Dependent Claims (99, 100, 101)
-
-
102. An apparatus for object detection comprising a computer system including a processor, a memory coupled with the processor, an input coupled with the processor for receiving images, and an output coupled with the processor for outputting information based on an object estimation, wherein the computer system further comprises means, residing in its processor and memory, for:
-
receiving images of an area occupied by at least one object; extracting image features including wavelet features from the images; classifying the image features to produce object class confidence data, wherein the classifying means use at least two sub-classifiers; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the means for performing data fusion comprises the means for; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown; wherein means for extracting image features comprises a means for; extracting wavelet coefficients of the at least one object in the images; and wherein the means for classifying the image features comprises processing the wavelet coefficients with at least one classification algorithm to produce object class confidence data; and wherein the means for extracting image features further comprises means for; detecting edges of the at least one object within the images; masking the edges with a background mask to find important edges; calculating edge pixels from the important edges; and producing edge density maps from the important edges, the edge density map providing the image features, and wherein the means for classifying the image features processes the edge density map with at least one of the classification algorithms to produce object class confidence data. - View Dependent Claims (103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113)
-
-
114. An apparatus for object detection comprising a computer system including a processor, a memory coupled with the processor, an input coupled with the processor for receiving images, and an output coupled with the processor for outputting information based on an object estimation, wherein the computer system further comprises means, residing in its processor and memory, for:
-
receiving images of an area occupied by at least one object; extracting image features including wavelet features from the images; classifying the image features to produce object class confidence data, wherein the classifying means use at least two sub-classifiers; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the means for performing data fusion comprises the means for; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown; wherein means for extracting image features comprises a means for; extracting wavelet coefficients of the at least one object in the images; and wherein the means for classifying the image features comprises processing the wavelet coefficients with at least one classification algorithm to produce object class confidence data; and wherein the means for extracting image features further comprises means for; receiving a stereoscopic pair of images of an area occupied by at least one object; detecting pattern regions and non-pattern regions within each of the pair of images using a texture filter; generating an initial estimate of spatial disparities between the pattern regions within each of the pair of images; using the initial estimate to generate a subsequent estimate of the spatial disparities between the non-pattern regions based on the spatial disparities between the pattern regions using disparity constraints; and
wherein the disparity constraints comprise consistency requirement constraints and local smoothness constraints, wherein the consistency requirement constraints are defined by - View Dependent Claims (115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125)
-
-
126. An apparatus for object detection comprising a computer system including a processor, a memory coupled with the processor, an input coupled with the processor for receiving images, and an output coupled with the processor for outputting information based on an object estimation, wherein the computer system further comprises means, residing in its processor and memory, for:
-
receiving images of an area occupied by at least one object; extracting image features including wavelet features from the images; classifying the image features to produce object class confidence data, wherein the classifying means use at least two sub-classifiers; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the means for performing data fusion comprises the means for; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown; where the at least one object comprises a vehicle occupant and the area comprises a vehicle occupancy area, and further comprising a means for processing the detected object estimate to provide signals to vehicle systems; wherein the signals comprise airbag enable and disable signals; wherein the apparatus further comprises a means for capturing images from a sensor selected from a group consisting of CMOS vision sensors and CCD vision sensors; and wherein at least one of the classification algorithms is selected from the group consisting of a trained C5 decision tree, a trained Nonlinear Discriminant Analysis network, and a trained Fuzzy Aggregation Network.
-
-
127. An apparatus for object detection comprising a computer system including a processor, a memory coupled with the processor, an input coupled with the processor for receiving images, and an output coupled with the processor for outputting information based on an object estimation, wherein the computer system further comprises means, residing in its processor and memory, for:
-
receiving a stereoscopic pair of images of an area occupied by at least one object; extracting image features from the images, with at least a portion of the image features being extracted by means for; detecting pattern regions and non-pattern regions within each of the pair of images using a texture filter; generating an initial estimate of spatial disparities between the pattern regions within each of the pair of images; using the initial estimate to generate a subsequent estimate of the spatial disparities between the non-pattern regions based on the spatial disparities between the pattern regions using disparity constraints; iteratively using the subsequent estimate as the initial estimate in the means for using the initial estimate to generate a subsequent estimate in order to generate further subsequent estimates of the spatial disparities between the non-pattern regions based on the spatial disparities between the pattern regions using the disparity constraints until there is no change between the results of subsequent iterations, thereby generating a final estimate of the spatial disparities; generating a disparity map of the area occupied by at least one object from the final estimate of the spatial disparities; classifying the image features to produce object class confidence data, wherein the classifying means use at least two sub-classifiers and with at least a portion of the classifying being performed by processing the disparity map with at least one classification algorithm to produce object class confidence data; and performing data fusion on the object class confidence data to produce a detected object estimate, wherein the means for performing data fusion comprises the means for; initially training the sub-classifiers in a supervised way by using the image features as inputs to the sub-classifiers and by using correct decisions known a priori as outputs of the sub-classifiers; training a fusion classifier by using confidence values generated by the trained sub-classifiers as inputs to the fusion classifier and by using correct decisions known a priori as outputs of the fusion classifier; and using the trained sub-classifiers and trained fusion classifier to perform data fusion to produce a detected object estimate when the correct decisions are unknown. - View Dependent Claims (128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144)
-
Specification