DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claims 1-20 are pending.
Claim Rejections - 35 USC § 103
The following is a quotation of pre-AIA 35 U.S.C. 103(a) which forms the basis for all obviousness rejections set forth in this Office action:
(a) A patent may not be obtained though the invention is not identically disclosed or described as set forth in section 102 of this title, if the differences between the subject matter sought to be patented and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art to which said subject matter pertains. Patentability shall not be negatived by the manner in which the invention was made.
Claim(s) 1-5 and 8-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Watanabe et al (US20040190766A1) in view of Rohwer et al (US20250086958A1).
Regarding claim 1, Watanabe teaches an apparatus, comprising:
a classifier comprising first features described by first coordinates relative to a center anchored coordinate;
a processor; and a non-transitory computer-readable medium comprising instructions that, when executed by the processor, cause the apparatus to:
(Watanabe, The control processing unit 15, which may be a personal computer for example, is comprised of hardware and software", [0028]; "The processing is executed in the control processing unit 15 by means of a CPU and software installed in advance", [0046]; "a reference point O is set near the center of the object.", [0036]; "a vector (u, v) from the reference point to a feature point", [0054]; Rohwer, "backendNet to learn to recognize objects of a particular class", [0057]; Watanabe teaches an apparatus that processes feature coordinates relative to a center-anchored point but relies on rigid template matching. Rohwer teaches a robust machine-learning classifier that uses localized footprint features)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to incorporate the teachings of Rohwer into the system or method of Watanabe in order to replace rigid template matching with a flexible classifier, thereby improving the system's ability to accurately categorize the center-anchored feature. The combination of Watanabe and Rohwer also teaches other enhanced capabilities.
The combination of Watanabe and Rohwer further teaches:
determine the center anchored coordinate;
(Watanabe, "a reference point O is set near the center of the object.", [0036]; determining the center anchored coordinate by setting a reference point near the object's center)
determine the first coordinates based on the center anchored coordinate;
(Watanabe, "The image of the vector OP at the time when the model pattern is taught is represented by (u, v)", [0041]; "a vector (u, v) from the reference point to a feature point", [0054]; determining the first coordinates based on the center anchored coordinate by establishing a vector from the reference center point to the feature point)
transform the first coordinates to second coordinates based on a matrix; and
(Watanabe, "Matrix elements r1-r9 in the rotating matrix in formula (1) can be defined variously.", [0036]; "a vector (u, v) from the reference point to a feature point is transformed into a vector (u′, v′)", [0054]; transforming the first coordinates to second coordinates using a rotation/transformation matrix to account for changes in orientation)
generate a classification based on values at the second coordinates.
(Watanabe, "a pattern matching is performed using the i-th transformed model pattern.", [0052]; "a local maximum point having a similarity equal to or higher than a preset value is searched for from results of the pattern matching.", [0055]; Rohwer, "The backendNet output channels can then be used ... to associate with any particular object class.", [0055]; "The placement parameter values are supplied by the output of a placement neural network", [0060]; Watanabe teaches performing pattern matching on the transformed coordinates. Rohwer teaches generating a classification using a neural network on transformed footprint coordinates. Together Watanabe and Rohwer teach generating a classification based on the transformed values to accurately classify objects despite spatial/orientation variances)
Regarding claim 2, the combination of Watanabe and Rohwer teaches its/their respective base claim(s).
The combination further teaches the apparatus of claim 1, further comprising a sensor configured to provide real-time sensed data and where the first coordinates are transformed to the second coordinates based on the real-time sensed data.
(Watanabe, "The image capturing means 14, which may be the conventionally known one such as a CCD video camera, is connected to a control processing unit 15", [0028]; "an image capturing command is delivered... and a two dimensional image including an image of one or more objects 33 is acquired", [0031]; an apparatus having a camera/sensor to capture real-time image data of objects, where coordinate transformations are performed to match the orientation of the object in the newly acquired sensed data)
Regarding claims 3 and 16, the combination of Watanabe and Rohwer teaches its/their respective base claim(s).
The combination further teaches the apparatus of claim 2, where the second coordinates are a rotation of the first coordinates about the center anchored coordinate.
(Watanabe, "Symbol R denotes rotation around a straight line passing through the point O... and φ denotes rotation around a straight line obtained by rotating a straight line passing through the point O", [0036]; the transformed second coordinates are a rotation of the first coordinates about the center anchored coordinate (point O))
Regarding claims 4 and 18, the combination of Watanabe and Rohwer teaches its/their respective base claim(s).
The combination further teaches the apparatus of claim 1, where the second coordinates are a scaling of the first coordinates about the center anchored coordinate.
(Watanabe, "In formulae (12), there is a relation of s=z0/z2... It is an amount or a scale that represents how many times the image size is scaled up or down", [0045]; "u' = s(r1u + r2v) (13)", [0060]; the second coordinates are scaled from the first coordinates about the center using a scaling factor 's')
Regarding claim 5, the combination of Watanabe and Rohwer teaches its/their respective base claim(s).
The combination further teaches the apparatus of claim 4, where the scaling is symmetric in two dimensions.
(Watanabe, "u' = s(r1u + r2v) v' = s(r4u + r5v) (13)", [0060]; applying the same scaling factor 's' to both the u' and v' dimensions, which constitutes symmetric scaling in two dimensions)
Regarding claim 8, the combination of Watanabe and Rohwer teaches its/their respective base claim(s).
The combination further teaches the apparatus of claim 1, where the first coordinates, the second coordinates, and the center anchored coordinate are within an image.
(Watanabe, "coordinate values (u, v) of the local maximum points in the image plane are extracted", [0055]; "the image (u0, v0) of the point O", [0038]; the center reference point and the coordinates are located within the two-dimensional image plane)
Regarding claims 9 and 20, the combination of Watanabe and Rohwer teaches its/their respective base claim(s).
The combination further teaches the apparatus of claim 1, where the first coordinates, the second coordinates, and the center anchored coordinate are within an array of photosites, and the processor is an image signal processor.
(Watanabe, "The image capturing means 14, which may be the conventionally known one such as a CCD video camera, is connected to a control processing unit 15 for visual sensor", [0028]; "it is enough to shift the grayscale pattern in units of picture element", [0053]; Rohwer, "Examples of processing circuitry 130 include microprocessors, application processors, display controllers, auxiliary processors, one or more sensor hubs", [0034]; "third dimension a horizontal pixel coordinate", [0042]; Watanabe teaches a CCD camera (inherently an array of photosites) capturing picture elements processed by a visual sensor control unit. Rohwer teaches utilizing dedicated display/auxiliary processors for image pixels. Together Watanabe and Rohwer teach coordinates corresponding to a photosite/pixel array processed by a signal processor. Incorporating Rohwer's specific image processing circuitry into Watanabe's CCD sensor system ensures efficient execution of transformations on pixel arrays)
Regarding claim 10, the combination of Watanabe and Rohwer teaches a method, comprising: determining a center anchored coordinate within an image;
(Watanabe, "image processing is performed by software to obtain a two dimensional image", [0031]; "a reference point O is set near the center of the object.", [0036]; determining a center anchored reference coordinate within a two-dimensional image)
determining first coordinates within an image based on the center anchored coordinate;
(Watanabe, Rohwer, see comments on claim 1; Watanabe teaches defining a vector (u,v) from the center reference point to a feature point)
multiplying the first coordinates by a transformation matrix to obtain second coordinates within the image; and
(Watanabe, "Matrix elements r1-r9 in the rotating matrix in formula (1) can be defined variously.", [0036]; "formulae (13) in which the geometrical transformation is represented by an affine transformation and which are as follows: u' = s(r1u + r2v) v' = s(r4u + r5v)", [0060]; multiplying the first coordinates by elements of an affine/rotation transformation matrix to obtain transformed second coordinates)
generating a classification based on image values at the second coordinates.
(Watanabe, Rohwer, see comments on claim 1; Watanabe and Rohwer collectively teach outputting an object classification based on the image features at the transformed coordinates)
Regarding claim 11, the combination of Watanabe and Rohwer teaches its/their respective base claim(s).
The combination further teaches the method of claim 10, where the transformation matrix is a pre-defined rotation.
(Watanabe, "R is varied from −180 to +180 in increments of 10, θ is varied from −90 to +90 in increments of 10, and φ is varied from −10 to +10 in increments of 10", [0049]; generating the transformation matrix using pre-defined rotation increments)
Regarding claim 12, the combination of Watanabe and Rohwer teaches its/their respective base claim(s).
The combination further teaches the method of claim 10, where the transformation matrix is a pre-defined scaling.
(Watanabe, "s is varied from 0.9 to 1.1 in increments of 0.05", [0048]; generating the transformation matrix using pre-defined scaling increments)
Regarding claim 13, the combination of Watanabe and Rohwer teaches its/their respective base claim(s).
The combination further teaches the method of claim 12, where the pre-defined scaling is asymmetric.
(Rohwer, "A scale is a pair (sh, sw) of vertical and horizontal multiplicative factors", [0058]; Watanabe teaches scaling coordinates using a predefined scaling factor ([0060]). Rohwer teaches applying separate vertical and horizontal scaling factors to patches. Incorporating Rohwer's distinct horizontal and vertical scaling factors into Watanabe's transformation would be obvious to account for asymmetric or skewed distortions in the captured image)
Regarding claims 14 and 17, the combination of Watanabe and Rohwer teaches its/their respective base claim(s).
The combination further teaches the method of claim 10, where the second coordinates are not a translation of the first coordinates.
(Watanabe, "In other words, it is assumed that x2=0 and y2=0. Thus, formulae (11) are replaced by formulae (12)", [0044]; excluding parallel displacement (translation) from the coordinate transformation equations, meaning the second coordinates are rotated or scaled but not translated relative to the first)
Regarding claim 15, the combination of Watanabe and Rohwer teaches an apparatus, comprising:
a machine learning logic configured to generate a classification for a detection window based on first coordinates that are offset from a center anchored coordinate;
(Watanabe, "a vector (u, v) from the reference point to a feature point", [0054]; Rohwer, "a placement neural network configured to process a patch of image data ... for aligning a footprint in the patch", [0008]; "A footprint is defined by a list of coordinates relative to the interest point.", [0059]; "backendNet to learn to recognize objects of a particular class", [0057]; Watanabe teaches establishing feature coordinates offset from a center reference point. Rohwer teaches a machine learning logic (e.g., backendNet) generating a classification for a detection window/patch based on offset footprint coordinates. Combining Watanabe and Rohwer would improve localized object recognition)
a processor; and a non-transitory computer-readable medium comprising instructions that, when executed by the processor, cause the apparatus to:
multiply the first coordinates by a transformation matrix to obtain second coordinates; and
(Watanabe, Rohwer, see comments on claim 10; Watanabe teaches multiplying the coordinates by an affine transformation matrix)
generate the classification based on values at the second coordinates.
(Watanabe, Rohwer, see comments on claim 1; Watanabe and Rohwer collectively teach generating the object classification based on the transformed coordinates)
Regarding claim 19, the combination of Watanabe and Rohwer teaches its/their respective base claim(s).
The combination further teaches the apparatus of claim 15, where the second coordinates correspond to pixels of an image.
(Watanabe, "it is enough to shift the grayscale pattern in units of picture element such that the picture element (u, v) in the original pattern is shifted to the picture element (u′, v′) in the transformed pattern", [0053]; the first and second coordinates correspond directly to the picture elements (pixels) of the image)
Claim(s) 6 is/are rejected under 35 U.S.C. 103 as being unpatentable over Watanabe et al (US20040190766A1) in view of Rohwer et al (US20250086958A1) and further in view of Tariq et al (US20190392268A1).
Regarding claim 6, the combination of Watanabe and Rohwer teaches its/their respective base claim(s).
The combination further teaches the apparatus of claim 1, where the first coordinates are retrieved from a pre-defined stride data structure according to a real-time budget constraint.
(Rohwer, "To support the grid-based approach, pointify module 202 transforms an input image into patches 230 of a given edge dimensions (“stride”).", [0079]; [0079]; Tariq, "The techniques discussed herein improve computer vision by increasing the accuracy of object detection and decreasing the compute time for obtaining accurate object identifications so that objects may be detected in real time", [0021]; Rohwer teaches transforming images and extracting coordinates using a predefined stride structure; Tariq teaches optimizing compute time limits for real-time object detection applications; together Rohwer and Tariq teach retrieving coordinates from a pre-defined stride data structure according to a real-time budget constraint)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the stride-based patch extraction of Rohwer with the real-time compute constraints of Tariq in order to ensure the object detection models execute within strict real-time deadlines for time-critical applications. The combination of Watanabe and Rohwer also teaches other enhanced capabilities.
Claim(s) 7 is/are rejected under 35 U.S.C. 103 as being unpatentable over Watanabe et al (US20040190766A1) in view of Rohwer et al (US20250086958A1) and further in view of Lelowicz et al (US20220335650A1).
Regarding claim 7, the combination of Watanabe and Rohwer teaches its/their respective base claim(s).
The combination does not expressly disclose but Lelowicz teaches the apparatus of claim 1, where the second coordinates are a mirroring or a flipping of the first coordinates about the center anchored coordinate.
(Lelowicz, "the image processing comprises at least one of a rotation, a scaling, a flipping, and an embedding in another image", [0009]; Lelowicz teaches applying a flipping transformation to image regions. Watanabe and Rohwer teach transforming coordinates about a center anchored coordinate)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to the flipping transformation taught by Lelowicz into the coordinate transformation techniques of Watanabe and Rohwer in order to improve the machine learning system's training and detection robustness against spatial and orientation variations, which predictably results in the second coordinates being a flipping of the first coordinates about the center anchored coordinate. The combination of Watanabe and Rohwer also teaches other enhanced capabilities.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JIANXUN YANG whose telephone number is (571)272-9874. The examiner can normally be reached on MON-FRI: 8AM-5PM Pacific Time.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Amandeep Saini can be reached on (571)272-3382. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272- 1000.
/JIANXUN YANG/
Primary Examiner, Art Unit 2662 7/7/2026
Prosecution Insights