DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Priority
The present application claims benefit of provisional application filed on 12/14/2023.
Information Disclosure Statement
The information disclosure statement(s) (IDS) submitted on 06/09/2025 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Objections
Claims 4-5 are objected to because of informalities in claim language. Each claim recites “The apparatus of claim [3/1], perform a third search,” which fails to properly provide a grammatical connection between the third search operation and the apparatus or instructions recited in claim 1. Correction is required, such as replacing “perform” with “wherein the instructions further cause the apparatus to perform,” or other equivalent language.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim 8-20 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 8 recites “performing a second at the second location.” It is unclear what second action is being performed at the second location because a verb or sequence of events is not being claimed after the term “a second”. Accordingly, claim 8 is rendered indefinite and rejected under 35 U.S.C. 112(b). Claims 9-14 are rejected under 35 U.S.C. 112(b) for being dependent on claim 8. For further 101 and 102/103 analysis, the limitation will be read as “performing a second search at the second location”.
Claims 15 recites “A machine learning logic comprising a plurality of detectors trained to” obtain a first location, search the first location, and determine a second location based on the first detection result. It is unclear whether the trained detectors themselves perform the recited search location management functions or whether separate search management logic performs those functions while the detectors perform object detection. The specification states that “The search management logic handles the control path for searching” (¶ [0219]) and describes the search management logic as obtaining search coordinates, initiating searches, assessing search results, and refining subsequent search coordinates, while separately describing the trained classifiers and detectors performing object detection within the search data path (see FIG. 12, Search Management, and Classification sections). Thus, the metes and bounds of the claim are unclear because it is uncertain which component performs the recited location-selection functions. Accordingly, claim 15 is rendered indefinite and rejected under 35 U.S.C. 112(b).
Claims 16-20 are rejected under 35 U.S.C. 112(b) as being indefinite for depending upon indefinite claim 15.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claim(s) 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The limitations, under their broadest reasonable interpretation, cover mental process (concept performed in a human mind, including as observation, evaluation, judgment, opinion, organizing human activity and mathematical concepts and calculations). The claim(s) recite(s) steps for determining a first and second location in an image to search based on a likelihood of a first object detection and a scan pattern, and performing subsequent first and second searches at respective locations. This judicial exception is not integrated into a practical application because the steps do not add meaningful limitations to be considered specifically applied to a particular technological problem to be solved. The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the steps of the claimed invention can be done mentally and no additional features in the claims would preclude them from being performed as such except for the generic computer elements at high level of generality (e.g. processor, memory, operating system, etc.).
According to the USPTO guidelines, a claim is directed to non-statutory subject matter if:
STEP 1: the claim does not fall within one of the four statutory categories of invention (process, machine, manufacture or composition of matter), or
STEP 2: the claim recites a judicial exception, e.g. an abstract idea, without reciting additional elements that amount to significantly more than the judicial exception as determined using the following analysis:
STEP 2A (PRONG 1): Does the claim recite an abstract idea, law of nature, or natural phenomenon?
STEP 2A (PRONG 2): Does the claim recite additional elements that integrate the judicial exception into a practical application?
STEP 2B: Does the claim recite additional elements that amount to significantly more than the judicial exception?
Using the two-step inquiry, it is clear that independent claims 8 and 15 are directed to an abstract idea (e.g. mental process) as shown below:
► STEP 1: Do the claims fall within one of the statutory categories?
Regarding claim 8, YES. Claim 8 is directed to a method (i.e. process) for determining two locations in an image to search based on a likelihood of an object detection and a scan pattern, and performing subsequent searches at respective locations.
Regarding claim 15, NO. Claim 15 is directed to “A machine learning logic” without reciting that the logic is embodied in a machine, manufacture, process, or composition of matter. The specifications do not expressly define “machine learning logic” as requiring physical structure. The closest corresponding disclosure describes classification or matching logic comprising trained classifiers and detectors, but expressly allows such logic to be implemented as logical threads, virtual machines, computer readable instructions, physical hardware, or combinations thereof (see FIG. 2, ¶ [0058], and ¶¶ [0240]-[0242] of the instant specifications). Accordingly, under the broadest reasonable interpretation, the claimed machine learning logic encompasses software or logic and therefore does not necessarily fall within one of the four statutory categories. Claims 16-20 fail to introduce physical hardware components for implementing the machine learning logic of independent claim 15. Therefore, claims 15-20 fail STEP 1 and are rejected under 35 U.S.C. 101 because the claimed “machine learning logic” does not fall within one of the four statutory categories of process, machine, manufacture, or composition of matter.
Due to claims 15-20 failing STEP 1, the remainder of the 101 analysis is directed to claim 8 and its dependent claims (claims 9-14).
► STEP 2A (PRONG 1): Is the claim directed to a law of nature, a natural phenomenon or an abstract idea?
Yes, the claim 8 is directed toward a mental process and/or mathematical concepts (i.e. abstract idea).
With regard to STEP 2A (PRONG 1), the guidelines provide three groupings of subject
matter that are considered abstract ideas:
Mathematical concepts - mathematical relationships, mathematical formulas or
equations, mathematical calculations;
Certain methods of organizing human activity - fundamental economic principles or practices (including hedging, insurance, mitigating risk); commercial or legal interactions (including agreements in the form of contracts; legal obligations; advertising, marketing or sales activities or behaviors; business relations); managing personal behavior or relationships or interactions between people (including social activities, teaching, and following rules or instructions); and
Mental processes - concepts that are practicably performed in the human mind (including an observation, evaluation, judgment, opinion).
Claim 8 comprises a mental process that can be practicably performed in the human mind (or generic computers or components configured to perform the method) and, therefore, an abstract idea.
Regarding independent claim 8, the method recites the mental steps of:
determining a first location to search in an image based on a first object detection
likelihood; (Mental processing including observation, judgement, and evaluation that can be performed mentally in the human mind by observing image data and making a judgement based on an evaluation of the observed image data)
performing a first search at the first location; (Mental processing by observing image data and evaluating the image data)
determining a second location to search in the image based on a scan pattern; (Mental processing by evaluating observed image data to make a judgment on second location)
and performing a second [search] at the second location (Mental processing by observing image data and evaluating the image data).
These limitations, as drafted, is a simple process that, under their broadest reasonable interpretation, covers performance of the limitations in the mind or by a human. The Examiner notes that under MPEP 2106.04(a)(2)(III), the courts consider a mental process (thinking) that "can be performed in the human mind, or by a human using a pen and paper" to be an abstract idea. CyberSource Corp. v. Retail Decisions, Inc., 654 F.3d 1366, 1372, 99 USPQ2d 1690, 1695 (Fed. Cir. 2011). As the Federal Circuit explained, "methods which can be performed mentally, or which are the equivalent of human mental work, are unpatentable abstract ideas the 'basic tools of scientific and technological work' that are open to all."' 654 F.3d at 1371, 99 USPQ2d at 1694 (citing Gottschalk v. Benson, 409 U.S. 63, 175 USPQ 673 (1972)). See also Mayo Collaborative Servs. v. Prometheus Labs. Inc., 566 U.S. 66, 71, 101 USPQ2d 1961, 1965 ('"[M]ental processes[] and abstract intellectual concepts are not patentable, as they are the basic tools of scientific and technological work'" (quoting Benson, 409 U.S. at 67, 175 USPQ at 675)); Parker v. Flook, 437 U.S. 584,589, 198 USPQ 193, 197 (1978) (same).
As such, a person could know or have a mental understanding of the likelihood of a first object being detected in an image and determine a first location to first search, perform the first search by observing the first location, mentally determine a second location to search in the image by observing the image in a pattern-like scan with their eyes (i.e. observe), and perform a second search at the second location by observing and analyzing the second location either mentally or using a pen and paper. The claim does not introduce any device or components, but it is noted, the mere nominal recitation of various steps being executed by a device/in a device (e.g. processing unit) would not take the limitations out of the mental process grouping. Thus, the claims recite a mental process.
STEP 2A (PRONG 2): Does the claim recite additional elements that integrate the judicial exception into a practical application?
NO, the claims do not recite additional elements that integrate the judicial exception into a practical application.
With regard to STEP 2A (prong 2), whether the claim recites additional elements that integrate the judicial exception into a practical application, the guidelines provide the following exemplary considerations that are indicative that an additional element (or combination of elements) may have integrated the judicial exception into a practical application:
an additional element reflects an improvement in the functioning of a computer, or an improvement to other technology or technical field;
an additional element that applies or uses a judicial exception to affect a particular treatment or prophylaxis for a disease or medical condition;
an additional element implements a judicial exception with, or uses a judicial exception in conjunction with, a particular machine or manufacture that is integral to the claim;
an additional element effects a transformation or reduction of a particular article to a different state or thing; and
an additional element applies or uses the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment, such that the claim as a whole is more than a drafting effort designed to monopolize the exception.
While the guidelines further state that the exemplary considerations are not an exhaustive list and that there may be other examples of integrating the exception into a practical application, the guidelines also list examples in which a judicial exception has not been integrated into a practical application:
an additional element merely recites the words "apply it" (or an equivalent) with the judicial exception, or merely includes instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea;
an additional element adds insignificant extra-solution activity to the judicial exception;
an additional element does no more than generally link the use of a judicial exception to a particular technological environment or field of use.
Independent claim(s) 8 does not recite any of the exemplary considerations that are indicative of an abstract idea having been integrated into a practical application. Claim 8 doesn’t disclose any computer components and although claim 8 recites an image and object detection, the claim merely uses these as the environment in which the abstract idea is performed. The claim does not recite any unconventional detector configuration, specialized hardware, improved computer architecture, or other technological improvement beyond the abstract idea itself. Rather, the claim broadly refers to conventional object detection performing the recited determination and search operations.
Although the specification discusses improvements such as reduced power consumption, reduced search time, adaptive search control, and detector optimization, these features and/or features that enable these improvements are not recited in claim 8. These features must be present in the claim, not the enclosed embodiments, in order to be eligible under 101. Thus, 8 merely recites determining successive search locations based on object detection information without reciting the technological improvements described in the specification.
Accordingly, claim 8 does not amount to significantly more than the judicial exception and is therefore directed to patent ineligible subject matter under STEP 2A (PRONG 2)
STEP 2B: Does the claim recite additional elements that amount to significantly more than the judicial exception?
NO, the claims do not recite additional elements that amount to significantly more than the judicial exception.
With regard to STEP 2B, whether the claims recite additional elements that provide significantly more than the recited judicial exception, the guidelines specify that the pre-guideline procedure is still in effect. Specifically, that examiners should continue to consider whether an additional element or combination of elements:
adds a specific limitation or combination of limitations that are not well-understood, routine, conventional activity in the field, which is indicative that an inventive concept may be present; or
simply appends well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception, which is indicative that an inventive concept may not be present.
Independent claim(s) 8 do not recite any additional elements that are not well understood, routine or conventional. The additional elements, individually and in combination, amount to not more than applying the abstract idea in a generic object-detection environment. The claim does not recite an unconventional detector, specialized hardware, other technological implementation that amounts to significantly more than the judicial exception. Accordingly, claim 8 does not include an inventive concept sufficient to transform an inventive concept into patent eligible subject matter.
Claims 9-14 do not integrate the mental process into a practical application or add significantly more to the mental process.
Regarding claim 9, the claim adds the limitation of capturing an image via forward-facing camera (generic device), and determining the first object detection likelihood based on a gaze point captured via an eye-tracking camera (generic device), this amounts to a mental process of observation, evaluation, and judgement (detecting objects) using generic devices performing well understood routine conventional activity (capturing images) and fails to remedy the abstract idea of claim 8.
Regarding claim 10, the claim adds the limitation of describing the scan pattern for searching in an image, this is further data gathering techniques utilizing observation and fails to remedy the abstract idea of claim 9.
Regarding claim 11, the claim adds further limitations to the scan pattern, thereby further providing observation techniques and failing to remedy the abstract idea of claim 10.
Regarding claim 12, the claim adds limitations classifying the first location, thereby providing judgments based on observations and evaluations, and failing to remedy the abstract idea of claim 11.
Regarding claim 13, the claim adds limitations further classifying the first location, thereby providing judgments based on observations and evaluations, and failing to remedy the abstract idea of claim 11.
Regarding claim 14, the claim adds limitations classifying the second location, thereby providing judgements based on observations and evaluations, and failing to remedy the abstract idea of claim 11.
Thus, since claim(s) 8-20 are: (a) directed toward an abstract idea, (b) do not recite additional elements that integrate the judicial exception into a practical application, and (c) do not recite additional elements that amount to significantly more than the judicial exception, it is clear that Claim(s) 8-20 are not eligible subject matter under 35 U.S.C 101.
Claim(s) 1-7 are also directed to an abstract idea even though the claims recite additional elements. These elements (apparatus, processor, computer-readable medium) do not integrate the judicial exception into a practical application and do not add significantly more than the judicial exception. Therefore, 1-7 are also rejected under 35 U.S.C. 101.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claim(s) 1, 3-4, 8, and 15-16 are rejected under 35 U.S.C. 103 as being unpatentable over Schneiderman (US 20040066966 A1) in view of Gualdi et al. (“Multistage Particle Windows for Fast and Accurate Object Detection”; copy provided by Examiner).
Regarding claim 1,
Schneiderman teaches: An apparatus (Schneiderman teaches a computer based object-finder system, including “computer 22” executing object-finder software that identifies detected object locations and orientations (Abstract; ¶¶ [0044]-[0045]).), comprising:
a processor (Schneiderman teaches the object finder software executes “upon execution by a processor of the computer 22” (¶ [0045]; also see ¶ [0073])); and a non-transitory computer-readable medium comprising instructions that, when executed by the processor, cause the apparatus to (¶ [0073] “software code may be stored as a series of instructions or commands on a computer-readable medium, such as a random access memory (RAM), a read only memory (ROM), a magnetic medium such as a hard-drive or a floppy disk, or an optical medium such as a CD-ROM”):
obtain an image (¶ [0044] “A digital image 16 is a typical input to the object finder”)
determine a first location to search in the image (Schneiderman teaches the object finder determines candidate image window locations and places “the image window 32 at a first of the set of image window locations to be evaluated” (see ¶¶ [0172]-[0175]; ¶ [0183]).);
perform a first search at the first location via a first detector (Schneiderman teaches the first sub-classifier evaluates the image window at the first candidate location and produces a corresponding log-likelihood value, and further teaches “evaluating sub-classifier
f
k
at the specified candidate locations” and placing the image window at the first location (¶¶ [0182]-0183]). Under the broadest reasonable interpretation, the first sub-classifier taught by Schneiderman, which evaluates a candidate image window to determine whether the object is present, corresponds to the claimed first detector.);
perform a second search at a second location via the first detector (“The image window 32 can then be shifted to a second location (as shown in FIG. 19)” and the same sub-classifier is evaluated at that location, Schneiderman goes on to explain “and the log-likelihood ratio for the sub-classifier at the second location can be found” (¶ [0183]).),
where the second location has a second object detection likelihood (As stated above, Schneiderman teaches application of the sub-classifier at the second location produces “the log-likelihood ratio for the sub-classifier at the second location” (¶ [0183]).);
Schneiderman fails to explicitly disclose: [determining] a first location to search in the image based on a first object detection likelihood; and where the first object detection likelihood is greater than the second object detection likelihood.
In a related art, Gualdi teaches: determine a first location to search in the image based on a first object detection likelihood (Gualdi teaches selecting particle window locations using detector derived likelihood information (The search “focus[es], in an iterative manner, on the exploration of the image toward the area where the target objects are more likely to be found” and the classifier response is used to “increasingly draw samples on the areas where the objects are potentially present and avoiding to waste search time over other regions. We call these samples ‘particle windows’ (PWs)” (Gualdi p. 1590 left column second full paragraph); The particle window state X = (
w
x
,
w
y
,
w
s
) identifies a window location and scale, and “the state pdf can be assumed proportional to the measurement likelihood function” (see p. 1593 left column section 3 and p. 1593 right column paragraph 1 of subsection 3.2 through end of paragraph on p. 1594); The resulting proposal distribution “will drive sampling of the next stage more toward portions of the state space where the classifier yielded high responses” (p. 1594 right column last paragraph though end of paragraph on p. 1595).) Under the broadest reasonable interpretation, selecting a particle window location according to detector-derived measurement likelihood corresponds to determining a location to search based on an object detection likelihood.).
Although Gualdi teaches selecting subsequent particle-window locations using progressively refined detector responses, such that “the samples at subsequent stages…concentrate more and more on the peaks of the distribution, i.e., where the response of the classifier is higher” (Gualdi p. 1595 left column first full paragraph), and Schneiderman teaches successive detector evaluations at first and second locations, neither Gualdi nor Schneiderman explicitly disclose evaluating candidate locations such that a first searched location has a greater object detection likelihood than a second searched location, as presently claimed. Rather, Gualdi teaches using object detection likelihood to guide the selection of subsequent search locations, not the particular ordering of candidate evaluations recited in the claim.
Gualdi is an analogous reference to Schneiderman because both are directed to computer implemented object detection in which a classifier evaluates candidate image window locations to determine whether an object is present.
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify the candidate location selection taught by Schneiderman to use Gualdi’s likelihood guided search strategy because Gualdi teaches “a data-driven and focused search” that directs the search “toward the area where the target objects are more likely to be found” while “avoiding to waste search time over other regions” (Gualdi p. 1590 left column second full paragraph). Gualdi further teaches that reducing the number of evaluated windows provides “a lower computational burden” (Gualdi Abstract). Thus, applying Gualdi’s likelihood guided selection to the object detector taught by Schneiderman would have predictably reduced unnecessary detector evaluations and computational burden while maintaining effective object detection.
Although Gualdi does not expressly disclose this particular ordering of candidate evaluations, it further would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to prioritize evaluation of candidate locations according to their object detection likelihoods, such that locations having greater object detection likelihoods are searched before locations having lower objected detection likelihoods. Gualdi teaches using detector responses to guide the search toward regions where objects are more likely to be found, and applying that teaching to the candidate location evaluation taught by Schneiderman would have predictably ordered the searches from higher to lower likelihood locations, thereby further reducing unnecessary detector evaluations and computational burden.
Regarding claim 3,
Schneiderman and Gualdi teach the apparatus of claim 1.
Gualdi further teaches: where the second search is performed in response to the first search having a first soft information level (Guadi teaches that a first search produces a classifier response, explaining that “the higher the degree of response R(w) is, the further w reached the end of the cascade, the more similar it is to the object model” (p. 1593 right column lines 6-8). Under the broadest reasonable interpretation and consistent with paragraph [0103] of the specification, which describes “soft information” as referring to information as having some “some uncertainty, probability, or likelihood”, the classifier response taught by Gaudi corresponds to the claimed first soft information level because it represents a measure of the likelihood that an evaluated location corresponds to the object model, thereby constituting information indicative of uncertain probability of likelihood associated with object detection. Gaudi further teaches using the classifier response to guide subsequent search by explaining that “the response (Rpw) of the specific classifier…to the particle window pw is exploited to determine the weight,” such that “the proposal distribution…will drive the sampling of the next stage more toward portions of the state space where the classifier yielded high responses” (p. 1594 right column last paragraph). Accordingly, “the samples at subsequent stages decrease in number but concentrate more and more on the peaks of the distribution, i.e., where the response of the classifier is higher” (p. 1595 left column first full paragraph). Thus, because Gualdi uses the classifier response generated during the first search to guide the proposal distribution that selects locations for the subsequent search, Gualdi teaches performing the second search in response to the first search having the claimed first soft information.).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify the search process of Schneiderman, as previously modified by Gualdi, such that the second search is performed in response to the first search having a first soft information level, as taught by Gualdi, because Gualdi teaches using the classifier response produced during the first search to guide the proposal distribution from which locations are selected during the subsequent search. Therefore, doing so would have predictably concentrated subsequent detector evaluations in regions more likely to contain the object while reducing unnecessary image window evaluations and computational burden.
Regarding claim 4,
Schneiderman and Gualdi teach the apparatus of claim 3.
Schneiderman and Gualdi further teaches or renders obvious: perform a third search at a third location via a second detector different than the first detector and where the third search is performed in response to the second search having a second soft information level greater than the first soft information level (Gualdi teaches performing a third search at a third location in response to a higher classifier response by explaining that “the proposal distribution…will drive the sampling of the next stage more toward portions of the state space where the classifier yielded high responses” (Gualdi p. 1594 right column last paragraph). Gualdi further teaches that “only the best responses will be held in account, while the others will be inhibited,” such that “the samples at subsequent stages decrease in number but concentrate more and more on the peaks of the distribution, i.e., where the response of the classifier is higher” (p. 1595 left column lines 1-19). As described in the instant application’s specification (¶ [0104]), soft information represents information indicative of uncertainty, probability, or likelihood associated with object detection. Accordingly, Gualdi’s higher classifier response corresponds to the claimed second soft information level greater than the first soft information level.
Schneiderman further teaches that “the object finder 18 uses a view-based approach with multiple classifiers that are each specialized to a specific orientation of the object,” and that “a different view-based classifier may be specialized to detect frontal views” (Schneiderman ¶ [0053]; ¶ [0055]).).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to further modify the search process of Schneiderman, as modified above in view of Gualdi, such that the third search is performed at the third location via a second detector different from the first detector in response to the second search having a second soft information level greater than the first soft information level because Gualdi teaches progressively concentrating subsequent searches on location associated with higher classifier responses, while Schneiderman teaches using different view-based classifiers specialized for detecting different object orientations. Applying Schneiderman’s specialized detector to Gualdi’s progressively refined search would have predictably improved detection of the increasingly likely object while allowing detector for the object’s expected appearance to be performed only after higher object detection likelihood had been established, thereby improving detection accuracy and reducing unnecessary computational processing.
Regarding claim 8,
Schneiderman in view of Gualdi teaches or renders obvious the method steps corresponding to: determining a first location to search in an image based on a first object detection likelihood; performing a first search at the first location; [search] at the second location, for the reasons discussed above regarding the apparatus of claim 1 (For sake of brevity refer back to corresponding teachings of Schneiderman and Gualdi).
Claim 8 differs from claim 1 principally by expressly reciting “A method” and “determining a second location to search in the image based on a scan pattern.”
Schneiderman and Gualdi further teach or render obvious: A method [for which the image-based object detection model is performing] (While the framework of Schneiderman in view of Gualdi, detailed in claim 1, describes the hardware of an apparatus, the apparatus is also taught with respect to a method being performed. Scheiderman and Gualdi both also expressly describe their frameworks as methods (see Schneiderman ¶ [0003] and Gualdi’s Abstract));
determining a second location to search in the image based on a scan pattern; and performing a second at the second location (Gualdi teaches a multistage search in which locations for a subsequent search are selected according to a sampling arrangement that is progressively refined based on classifier feedback, explaining “multistage strategy where the proposal distribution is progressively refined by taking into account the feedback of the classifiers” (Abstract), and that the classifier response is used so that “the proposal distribution…will drive the sampling of the next stage more toward portions of the state space where the classifier yielded high responses” (p. 1594 right column last paragraph). Thus, the proposal distribution determines the particle window locations evaluated during the subsequent search stage. As described in the instant application specifications, a scan pattern can be “random, ML-driven, [or] randomized,” (see ¶ [0097] of instant specification). Accordingly, Guildi’s ML-driven, sampling-based technique for determining successive particle window locations corresponds to the claimed scan pattern. The resulting particle window selected according to the scan pattern corresponds to the second location, and evaluating the particle window with the classifier reasonably corresponds to performing the second search at the second locations.).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify the search process of Schneiderman, as modified above in view of Gualdi, to determine the second location based on Gualdi’s multistage sampling pattern and to perform the second search at that location because Gualdi teaches progressively refining the proposal distribution to direct subsequent classifier evaluations toward image regions more likely to contain the object. Doing so would have predictably reduced unnecessary image window evaluations and corresponding computational burdens, while retaining focused coverage of the image search area.
Regarding claim 15
Schneiderman in view of Gualdi teaches or renders obvious the operations corresponding to obtaining a first location having an object detection likelihood and searching the first location with a detector, for the reasons discussed above regarding claim 1. Schneiderman in view of Gualdi also teaches determining a subsequent location based on the result produced by a preceding detector search, for the reasons discussed in claim 3. For brevity, the corresponding teachings of Schneiderman and Gualdi set forth in claims 1 and 3 are incorporated herein.
Claim 15 differs principally by reciting those operations as being performed by machine learning logic comprising a plurality of detectors trained to perform the recited operations.
Schneiderman and Gualdi further teach or render obvious:
A machine learning logic comprising a plurality of detectors trained to: obtain a first location having a first object detection likelihood; search the first location via a first detector to generate a first detection result; and determine a second location to search based on the first detection result (Schneiderman teaches object detection logic comprising a plurality of trained detectors, explaining “the object finder 18 uses a view-based approach with multiple detectors or classifiers (the terms "detectors" and "classifiers" are used interchangeably hereinbelow) that are each specialized to a specific orientation” and that “each classifier is trained to detect…one orientation of a particular object” (Schneiderman ¶ [0053]). Schneiderman further explains each classifier is specialized to detect the object orientation for which the classifier was trained and that the classifiers are applied to an input image to detect object locations and generate corresponding detection results (Schneiderman ¶¶ [0053]-[0056]).
Gualdi teaches using the result generated by a classifier search to determine a subsequent search location and resulting proposal distribution drives sampling in the next stage toward portions of the areawhere the classifier produced higher responses (see Guildi p. 1594 right column last paragraph through end of paragraph on p. 1595, including “the response…of the specific classifier…to the particle window pw is exploited to determine the weight,” and the resulting “proposal distribution…will drive sampling of the next stage more toward portions of the state space where the classifier yielded high responses”). Accordingly, the classifier response generated by searching the first location corresponds to the first detection result, and using that response to guide the next stage particle-window location corresponds to determining the second location based on the first detection result.).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to implement the likelihood-guided search process of Schneiderman, as modified above (in claims 1 and 3) in view of Gualdi, using Schneiderman’s object finder comprising a plurality of trained view-based detectors, such that a first trained detector searches the first location to generate a first detection result and the machine learning logic uses that result to determine the second location, as taught by Gualdi. Doing so would have predictably used Schneiderman’s known trained detector model to implement Gualdi’s result-responsive search process, thereby directing subsequent detector evaluations toward image locations more likely to contain the object while reducing unnecessary image window evaluations and computational burden.
Regarding claim 16
Schneiderman and Gualdi teach the machine learning logic of claim 15.
Gualdi further teaches: where the second location is further based on a scan pattern (As discussed previously discussed with respect to claims 3 and 8, Gualdi teaches refining the proposal distribution based upon the classifier response to determine subsequently evaluated image locations. For brevity, refer back to corresponding teachings found in claims 3 and 8. Thus, Gualdi teaches determining subsequent search locations according to a classifier responsive scan pattern.
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to determine the second location according to Gualdi’s adaptive scan pattern because doing so predictably concentrates subsequent searches in image regions more likely to contain the object while reducing unnecessary evaluations).
Claim(s) 2, 5-7, and 9 are rejected under 35 U.S.C. 103 as being unpatentable over Schneiderman (US 20040066966 A1) in view of Gualdi et al (“Multistage Particle Windows for Fast and Accurate Object Detection”; copy provided by Examiner), and in further view of Neven (US 20120290401 A1).
Regarding claim 2,
Schneiderman and Gualdi teach the apparatus of claim 1, a first object detection likelihood.
Schneiderman and Gualdi fail to explicitly disclose: a forward-facing camera and an eye-tracking camera and where the image is captured via the forward-facing camera and the first object detection likelihood is based on user interest determined via the eye-tracking camera.
In a related art, Neven teaches: a forward-facing camera and an eye-tracking camera and where the image is captured via the forward-facing camera and the first object detection likelihood is based on user interest determined via the eye-tracking camera (Neven teaches a forward-facing camera and an eye-tracking camera, stating “The device 100 may also include… a scene camera (e.g., video camera) 120, a gaze tracking camera 121,” “the scene camera 120 may be forward facing to capture at least a portion of the real-world view perceived by the user,” and “Gazing tracking camera 121 is positioned to acquire eye images (e.g., video images) of the user's eye. These eye images can then be analyzed to generate gaze direction information, which can be used to determine the user's gazing direction. The gaze direction information can be correlated to the scene images acquired by scene camera 120 to determine at what item (e.g., person, place, or thing) the user was directly looking at when viewing the external scene through lens elements 110 and 112 (¶¶ [0019]-[0021], also see FIGs. 1-2). Nevin further teaches “the eye image is analyzed to generate gaze direction information” and “the gazing information extracted from the eye images is used to determine which item within the external scene the user is staring or gazing directly at” (¶¶ [0037]-[0038]). Under the broadest reasonable interpretation of the claim, the gaze direction information identifying the item at which the user is looking corresponds to the claimed user interest, and the identified item corresponds to the claimed first object detection likelihood being based on user interest determined via the eye-tracking camera.).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to further modify the apparatus of Schneiderman, as modified by Gualdi, to include the forward-facing camera and eye-tracking camera taught by Neven, such that the first object detection likelihood is based on user interest determined via the eye-tracking camera because Gualdi teaches correlating gaze direction information with images captured by the forward-facing scene camera to determine the item at which the user is looking. Doing so would have predictably prioritized detector evaluations toward image locations that correspond to the user’s visual attention, thereby improving search efficiency while reducing unnecessary image window evaluations and associated computational burdens. All three references are directed detecting or recognizing objects within a captured image by evaluating localized image regions (e.g. sub-image windows or regions of interest).
Regarding claim 5,
Schneiderman and Gualdi teach the apparatus of claim 1.
Gualdi further teaches: perform a third search at a third location via the first detector, where the third location has a third likelihood (As detailed above, Gualdi teaches performing multiple subsequent searches in response to the classifier response produced during an earlier search, stating “the proposal distribution…will drive the sampling of the next stage more toward portions of the state space where the classifier yielded high responses” (p. 1594 right column last paragraph), and further teaches, “the samples at subsequent stages decrease in number but concentrate more and more on the peaks of the distribution, i.e., where the response of the classifier is higher” (p. 1595 left column first full paragraph), thereby teaching selecting multiple subsequent particle window locations from the proposal distribution for evaluation by the classifier. As detailed above and described in the instant application’s specification (¶ [0104]), soft information represents information indicative of uncertainty, probability, or likelihood associated with object detection. Accordingly, Gualdi’s classifier response corresponds to the claimed second soft information level, and the multiple particle window locations selected for the subsequent search reasonably corresponds to the claimed second and third locations having corresponding object-detection likelihoods).
Schneiderman and Gualdi fail to explicitly disclose: performing location search with respect to user interest
In a related art, Neven teaches: determining an image location associated with an item of user interest based on gaze information (¶ [0038] “a gazing direction can be inferred from the gazing information, which is then used to select a localized region within the corresponding captured scene image. This localized region can then be analyzed using other techniques to lock onto a particular item at which the user may be gazing at.”).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify the search process of Schneiderman, as previously modified by Gualdi, to use the gaze derived or image location taught by Neven as an additional basis for the subsequent search locations because Neven teaches identifying a localized scene region corresponding to the item viewed by the user, while Gualdi teaches selecting multiple subsequent particle window locations according to classifier responses. Using gaze derived user interest location to guide Gualdi’s subsequent searches would have predictably concentrated the second and third detector searches on multiple promising location associated with the object being viewed by the user, thereby reducing unnecessary image window evaluations while improving efficiency of user interest-based object detection.
Regarding claim 6,
Schneiderman, Gualdi, and Neven teaches the apparatus of claim 5.
Schneiderman, Gualdi, and Neven further teaches or renders obvious: where the second location and the third location are a first stride length from the first location (Schneiderman repeatedly searching image locations using a fixed detection window, stating “ input to the classifier 32 is fixed-size window sampled from an input image” and that “The object finder must apply the classifier 34 repeatedly to original image 58 for all possible (and, maybe overlapping) positions of this rectangular image window 32 (see Schneiderman ¶¶ [0051]-[0052] and corresponding FIGs. 4A and 8).
Gualdi further teaches expressly teaches that the spacing of search window locations is defined by stride, stating “the cardinality of the SWS also depends on the degree of coarseness for the scattering of the windows, i.e., pixel and scale strides” (p. 1593 right column last paragraph of 3.1), and further describes a multistage search in which “At the first level, a pixel stride of 6 is employed,” followed by, “pixel stride 3,” and then refinement by “fixing pixel stride equal to 1 in a 3 3 neighborhood” (p. 1592 right column second paragraph of 2.2). Thus, Gualdi teaches searching multiple neighboring window locations positioned according to a common pixel stride, corresponding to the second and third locations each being a first stride length from the first location.).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to implement the subsequent searches of Schneiderman, as modified above in view of Gualdi and Neven, such that the second location and the third location are each positioned a first stride length from the first location because Schneiderman teaches repeatedly applying a fixed detection window at shifted image positions, while Gualdi expressly teaches defining the displacement between neighboring search window positions using a pixel stride. Applying Gualdi’s known stride-based arrangement to Schneiderman’s repeated window searches would have predictably provided a consistent method for evaluating neighboring image locations while reducing unnecessary image window evaluations and computational burden.
Regarding claim 7,
Schneiderman, Gualdi, and Neven teach the apparatus of claim 6.
Schneiderman, Gualdi, and Nevin further teaches or renders obvious: where the first stride length is based on the first soft information level (As previously discussed, with regard to claim 6, Gualdi teaches positioning neighboring search window locations using a pixel stride (refer to Gualdi’s teachings found above in claim 6’s 103 rejection). Gualdi further teaches adapting the proposal distribution according to the classifier response obtained during the earlier search because “the response…of the specific classifier…to the particle window pw is exploited to determine the weight,” and the resulting “proposal distribution…will drive sampling of the next stage more toward portions of the state space where the classifier yielded high responses” (p. 1594 right column last paragraph through end of paragraph on p. 1595). Guadi further teaches that its statistical approach “does not employ a rigid grid structure in the refinement of the search, allowing the “radius” to be adapted based on the response of the classifier” (p. 1593 left column lines 5-7). As described above and in paragraph [0103] of the instant application’s specification, the classifier response corresponds to the claimed first soft information level. Accordingly, because Gualdi teaches both stride-based neighboring search locations and adapting subsequent searches according the classifier response, these teachings suggest or render obvious basing the claimed first stride length on the first soft information.).
Thus, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to further modify the stride-based search process of Schneiderman, as previously modified by Gualdi and Neven, such that the first stride length is based on the first soft information level because Gualdi teaches adapting the spatial amount of subsequent searches according to the classifier response produced during the earlier search (i.e. first soft information level). Applying Gualdi’s response adaptive search refinement to the stride-based neighboring search arrangement (discussed above in claim 6) would have predictably focused detector evaluations closer to locations associated with higher object detection likelihoods while reducing unnecessary image window evaluations and computational burdens.
Regarding claim 9,
Schneiderman and Gualdi teach the method of claim 8.
Schneiderman and Gualdi fail to explicitly disclose: capturing the image via a forward-facing camera, and determining the first object detection likelihood based on a gaze point captured via an eye-tracking camera.
In a related art, Neven teaches: capturing the image via a forward-facing camera, and determining the first object detection likelihood based on a gaze point captured via an eye-tracking camera (See Neven ¶¶ [0020]-[0021], ¶¶ [0035]-[0037], and further explanation provided in claim 2’s 103 rejection, found above.).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to further modify the method of Schneiderman, as modified by Gualdi, to capture the image via the forward-facing camera and determine the first object detection likelihood based on the gaze point captured via the eye-tracking camera, as taught by Neven, because Neven teaches using gaze direction information to identify the item within the captured scene that is the subject of the user’s visual attention. Doing so would have predictably focused object detection on image regions corresponding to the user’s gaze, thereby improving the search efficiency while reducing unnecessary image window evaluation and computational burdens.
Claim(s) 10-14 are rejected under 35 U.S.C. 103 as being unpatentable over Schneiderman (US 20040066966 A1) in view of Gualdi et al (“Multistage Particle Windows for Fast and Accurate Object Detection”; copy provided by Examiner), in further view of Neven (US 20120290401 A1), and in further view of Shustorovich et al. (US 20130182951 A1).
Regarding claim 10,
Schneiderman, Gualdi, and Neven teach the method of claim 9, and performing the detector searches according to a scan pattern, as seen in independent claim 8.
Schneiderman, Gualdi, and Neven fail to explicitly disclose: where the scan pattern comprises a center-to-outward spiral pattern.
In a related art, Shustorovich teaches: where the scan pattern comprises a (Shustorovich teaches progressively defining image test windows according to “a substantially spiral pattern,” wherein the spiral traversal proceeds from an outer image region toward an inner image region (Shustorovich ¶¶ [0144]-[0145]).).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify the scan pattern of Schneiderman, as previously modified in view of Gualdi and Neven, to use Shustorovich’s known spiral arrangement extending outward from the first location. Reversing the traversal direction of the known spiral scan pattern would have been a predictable implementation choice that merely changes the order in which the spiral-arranged image locations are evaluated while retaining the same spiral search pattern. Doing so would have predictably provided a systematic coverage of image locations surrounding the first location while reducing redundant or unnecessary image window evaluations.
Regarding claim 11,
Schneiderman, Gualdi, Neven, and Shustorovich teach the method of claim 10.
Gualdi further teaches where a stride size of the scan pattern is determined based on a detection result of the first search (Gualdi teaches that “NSW is a function of the image size, of the sliding steps (x,y) (also known as pixel strides), and…scale step s (also known as scale stride),” and that “the cardinality of the SWS also depends on the degree of coarseness for the scattering of the windows, i.e., pixel and scale strides.” (Gualdi p. 1589 right column second full paragraph; p. 1593 right column last paragraph of 3.1). Gualdi further teaches that “proposal distribution is progressively refined by taking into account the feedback of the classifiers,” thereby “allowing the “radius” to be adapted based on the response of the classifier” (Abstract; p. 1593 left column lines 4-7). Thus, Gualdi teaches that the classifier response adaptively controls the spatial coverage of the subsequent search, while the pixel and scale strides control the spacing of the subsequent search locations.
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to base the stride length of the subsequent scan pattern on the preceding detection result because Gualdi teaches adapting the spatial extent of the subsequent search based on the classifier response while using pixel and scale strides to control the spacing of the search locations. Using the classifier response to determine the stride length would have been a predictable implementation choice that adaptively refined the search while reducing unnecessary search evaluations.).
Regarding claim 12,
Schneiderman, Gualdi, Neven, and Shustorovich teach the method of claim 11.
Schneiderman further teaches: where the first search comprises classifying the first location with a first classifier tier and classifying the first location with a second classifier tier different than the first classifier tier (¶ [0013] “Each of the plurality of view-based classifiers includes a plurality of cascaded sub-classifiers. The cascaded stages may be arranged in ascending order of complexity and computation time.”; ¶ [0167] “the detector may be applied in a cascade of sequential stages of partial evaluation, where each stage performs a partial evaluation of the classifier, i.e., a subset of sub-classifiers. That is, each classifier is decomposed into a series of stages where each stage contains one or more sub-classifiers”; ¶ [0169] “the cascade evaluation strategy focuses on sub-classifier-by-sub-classifier analysis”; Accordingly, Schneiderman teaches or reasonably suggests successive classifier evaluations reasonably corresponding to the claimed first and second classifier tiers.).
Regarding claim 13,
Schneiderman, Gualdi, Neven, and Shustorovich teach the method of claim 12.
Schneiderman further teaches: where the first search further comprises classifying the first location with a third classifier tier different than the first classifier tier and the second classifier tier (¶ [0167] “the detector may be applied in a cascade of sequential stages of partial evaluation, where each stage performs a partial evaluation of the classifier, i.e., a subset of sub-classifiers. That is, each classifier is decomposed into a series of stages where each stage contains one or more sub-classifiers; ¶ [0169] “the cascade evaluation strategy focuses on sub-classifier-by-sub-classifier analysis and builds on earlier computations to generate the final result”; ¶ [0180] “Fi is the set of sub-classifiers associated with a particular stage… Each stage may have as few as one sub-classifier”; Accordingly, Schneiderman teaches or reasonably suggests continuing the successive classifier evaluations beyond the preceding evaluations, reasonably corresponding to the claimed third classifier tier.).
Regarding claim 14,
Schneiderman, Gualdi, Neven, and Shustorovich teach the method of claim 12.
Schneiderman further teaches: where the first search comprises classifying the second location with the first classifier tier and classifying the second location with the second classifier tier (¶ [0052] “The object finder must apply the classifier 34 repeatedly to original image 58 for all possible (and, maybe overlapping) positions of this rectangular image window 32”; ¶ [0168] “ The set of image window 32 locations to be evaluated can initially be a set of all possible image window 32 locations”; ¶ [0173] “ the object finder program 18 may evaluate a single stage… for each member of a set of image window 32 locations to be evaluated… the log-likelihood ratio generated by the application of a sub-classifier to a location.”; ¶ [0182] “At block 144 … a first sub-classifier… can be applied…This process repeats for all remaining sub-classifier in the stage.”; Accordingly, Schneiderman teaches or reasonably suggests performing the same successive classifier evaluations for each candidate image location, reasonably corresponding to classifying the recited second location using the claimed first and second classifier tiers).
Claim(s) 17-20 are rejected under 35 U.S.C. 103 as being unpatentable over Schneiderman (US 20040066966 A1) in view of Gualdi et al (“Multistage Particle Windows for Fast and Accurate Object Detection”; copy provided by Examiner), and in further view of Shustorovich et al. (US 20130182951 A1).
Regarding claim 17,
Schneiderman and Gualdi teach the machine learning logic of claim 16, including a scan pattern.
Schneiderman and Gualdi fail to explicitly disclose: where the scan pattern comprises a first center-to-outward spiral pattern centered about the first location.
In a related art, Shustorovich teaches: where the scan pattern comprises a first (Shustorovich teaches progressively defining image test windows according to “a substantially spiral pattern,” wherein the spiral traversal proceeds from an outer image region toward an inner image region (Shustorovich ¶¶ [0144]-[0145]).).
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify the scan pattern of Schneiderman, as previously modified in view of Gualdi and Neven, to use Shustorovich’s known spiral pattern because spiral traversal is a known scanning technique for systematically evaluating neighboring image locations. Although Shustorovich teaches traversing the spiral from the outer portion toward the center reversing the traversal to begin at the determined first location and proceed outward would have been a predictable variation that traverses the same spiral geometry in the opposite direction while prioritizing locations nearest to the determined first location.
Regarding claim 18,
Schneiderman, Gualdi, and Shustorovich teach the machine learning logic of claim 17.
Schneiderman in view of Gualdi further teaches or reasonably suggests: search the second location via the first detector to generate a second detection result; determine a third location to search based on the second detection result(Schneiderman ¶¶ [0051]-[0052] “The object finder must apply the classifier 34 repeatedly to original image 58 for all possible (and, maybe overlapping) positions of this rectangular image window 32”; ¶ [0173] “the object finder…may evaluate a single stage… for each member of a set of image window 32 locations to be evaluated… total log-likelihood, contains terms resulting from sub-classifiers already applied to the location.”; ¶ [0182] “At block 144 … a first sub-classifier… can be applied…This process repeats for all remaining sub-classifier in the stage.”; As discussed regarding claims 3, 5, and 15, Gualdi teaches using the classifier response obtained from a presently evaluated location to determine subsequently evaluated locations).
Shustorovich further teaches: where the third location is based on a second center-to-outward spiral pattern centered about the second location(Shustorovich teaches progressively defining image test windows according to “a substantially spiral pattern,” wherein the spiral traversal proceeds from an outer image region toward an inner image region (Shustorovich ¶¶ [0144]-[0145]). As discussed with respect to claim 17, although Shustorovich teaches traversing the spiral from the outer portion toward the center, reversing the traversal to proceed from the center outward would have been a predictable variation).
Thus, Schneiderman teaches searching successive image locations using the detector, Gualdi teaches iteratively determining each subsequent search location based on the preceding detection result, and Shustorovich teaches performing the search according to a spiral scan pattern. Together the references reasonably suggest determining the third location according to a second center-to-outward spiral pattern centered about the second location.
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to apply the same modified center-to-outward spiral scan pattern to each subsequently determined search location because Gualdi’s iterative search repeatedly determines subsequent search locations based on preceding detection results.
Regarding claim 19,
Schneiderman, Gualdi, and Shustorovich teach the machine learning logic of claim 16.
Schneiderman further teaches: where the first detector comprises at least a first classifier tier and a second classifier tier (¶ [0013] “Each of the plurality of view-based classifiers includes a plurality of cascaded sub-classifiers. The cascaded stages may be arranged in ascending order of complexity and computation time.”; ¶ [0128] “the detector may be applied in a cascade of sequential stages of partial evaluation, where each stage performs a partial evaluation of the classifier… each classifier is decomposed into a series of stages where each stage contains one or more sub-classifiers.”; ¶ [0180] “Fi is the set of sub-classifiers associated with a particular stage… Each stage may have as few as one sub-classifier”; Accordingly, Schneiderman teaches or reasonably suggests successive classifier evaluations reasonably corresponding to the claimed first and second classifier tiers.).
Regarding claim 20
Schneiderman, Gualdi, and Shustorovich teach the machine learning logic of claim 19.
Schneiderman further teaches: where the first detection result is based on a first result of the first classifier tier and a second result of the second classifier tier (¶ [0173] “the object finder.. may evaluate a single stage… the object finder 18 can keep, for each of the set of image window 32 locations to be evaluated, a partial calculation of equation …that may be referred to as a total log-likelihood… total log-likelihood, contains terms resulting from sub-classifiers already applied to the location”; ¶ [0182] “At block 144 .. a first sub-classifier… can be applied … This process repeats for all remaining sub-classifier in the stage.”; ¶ [0201] “the detection value (i.e., the left side of equations 8 and 8A) is computed for all viewpoints and at all locations”; Accordingly, Schneiderman teaches or reasonably suggests generating the detection result from the accumulated results of the successive classifier evaluations, reasonably corresponds to the claimed first and second classifier tier results).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SAMUEL DAVID BAYNES whose telephone number is (571)272-0607. The examiner can normally be reached Monday - Friday 8:00 am - 5:00 pm.
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, Stephen R Koziol can be reached at (408)918-7630. 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.
/SDB/
Samuel D. Baynes
Examiner, Art Unit 2665
/Stephen R Koziol/Supervisory Patent Examiner, Art Unit 2665
Prosecution Insights