DETAILED ACTION
This non-final office action is responsive to application 17/960,926 with applicant’s amendments and request for reconsideration as submitted 22 May 2026.
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim status at present stage of examination is as follows.
Pending: claims 1, 5, 7-14, 16-18 and 20-26
Amended: claims 1, 5, 8, 10-11 and 16-18
New: claims 21-26
Independent: claims 1, 10 and 21
Cancelled: claims 2-4, 6, 15 and 19
Response to Remarks
Applicant’s responsive remarks filed 05/22/26 are considered together with amendments in light of the remaining issues which are addressed as follows.
The rejection of claims 1-20 under 35 U.S.C. 101 as being directed to an abstract idea without significantly more is withdrawn in light of amendments, the remarks found at least partly persuasive, and upon further consideration of the amended claim such that eligibility is by practical application and/or significantly more satisfied by additional elements with consideration of the claim as a whole.
The rejection under 35 U.S.C. 103 as being obvious over a combination of prior arts is given updated search and consideration based on amendments of varying scope. Prior art discovery reveals newly applied reference Udutalapally as a significant finding, in addition to newly applied references Hadar and Harikumar. In view of the foregoing, an updated rejection is detailed below in a finding of obviousness. Applicant’s remarks regarding the prior art have been considered, but they are moot in view of the new grounds of rejection as necessitated by applicant’s amendments.
Claim Objections
Claims 9-10, 14 and 26 are objected to because of the following reasons:
Claims 9 and 14 limitation recites “data and time” should read “date and time”
Claim 10 limitation of generating notifications recites “comprising the least first type” should read “the at least first type” similarly as in claims 1 or 21.
Claim 26 is a substantial duplicate of claim 25, see MPEP 608.01(m)
Appropriate correction is required.
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.
Claims 10-14, 16-18 and 20 are rejected under 35 U.S.C. 112(b), as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor, regards as the invention. Particularly, limitation upon receiving appears to inadvertently amend by striking out the term “tag” leaving no verb in the phrase of limitation and creating an antecedent basis issue where it is subsequently recited in limitation correlate “the at least first type of subject and tagged source metadata.” The limitation should recite automatically tagging similar to claims 1 and 21 which will address the issue. Claims depending from 10 include the issue stemming from claim 10. Therefore, claims 10-14, 16-18 and 20 are rejected under 35 U.S.C. 112(b).
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.
Claims 1, 10 and 21 are rejected under 35 U.S.C. 103 as being unpatentable over:
Capota et al., US PG Pub No 2020/0410288A1 hereinafter Capota, in view of
Udutalapally et al., “sCrop: A Novel Device for Sustainable Automatic Disease Prediction, Crop Selection, and Irrigation in Internet-of-Agro-Things for Smart Agriculture” hereinafter Udutalapally, in view of
Hadar et al., US PG Pub No 2018/0300341 hereinafter Hadar, and further in view of
Harikumar et al., US PG Pub No 2022/0391633A1 hereinafter Harikumar.
Claim language is in bold with markup being applicant’s amendments.
With respect to claim 1, Capota teaches:
A method of expediting distributed feedback for training of supervised learning models to identify one or more types of subjects within images {Capota [0024] “techniques may be used for supervised learning, for example, classification” Figs 2-3 and 8 illustrate distributed environment for model management and training [0130,134] “train an image recognition model… object recognition”}, the method comprising:
in association with at least a first campaign for development of at least a first model for classifying at least a first type of subject under at least one of one or more defined classifying conditions, selecting a plurality of potential feedback source devices from a network of available feedback source devices based on identified lighting and/or weather conditions associated with each of the network of available feedback source devices as corresponding to the at least one of the one or more defined classifying conditions {Capota Fig 6:A110-120 shows Campaign and Selection of Devices, the campaign being for model development shown Fig 8 between server and devices, similar Figs 7, 3, [0039] “developing, tuning/optimizing, troubleshooting and managing models” models to classify e.g. [0134-35] “Classification examples include object recognition (traffic signs, objects in from of a vehicle, etc.), face recognition” cont’d “CNN… image classification” again at [0083,28]. Conditions are disclosed [0131] “light conditions… model needs to be able to recognize, for example, a traffic sign in rainy conditions with the same accuracy as when the sun is shining” and/or [0113,096] “weather conditions”};
Capota further discloses [0101,103] “push notifications from the campaign server… user defines the campaign through a label and any additional metadata”.
However, Capota does not appear to fairly disclose the image between device and network which is taught by Udutalapally:
generating notifications from a central computing network to each of the selected [[a]] plurality of potential feedback source devices requesting responsive images of the one or more types of subjects {Udutalapally [P.17528 Sect. V.A] “notify famers about the infected crop” shown at Fig 3-right “Real-time crop disease prediction notifications” bidirectional arrows note crop image for the sCrop mobile application on smartphone devices Fig 13 and in communication with cloud Fig 2, the subject including crop/leaf/plant Figs 4-7, Alg.1};
upon generating the notifications, establishing a respective feedback connection between each of the plurality of potential feedback source devices and a data storage network associated with the at least a first model {Udutalapally [P.1717-28 Pg.Brk] “5G, WiFi” as “sCrop device communicates with the upper layers in wireless” describes communication between the devices, shown Figs 2-3 and implemented Alg.1 [P.17531] “SendCloud…ReadCloudtoMobileApp” connects for feeding images read from camera e.g. [P.17529 ¶2] “feed in images of the crops to the Deep Learning (DL) model” and further shows data storage in cloud network at Fig 3};
Udutalapally is directed to training deep learning classifier models with distributed cloud environment thus being analogous. A person having ordinary skill in the art would have considered it obvious prior to the effective filing date to generate notifications over cloud using sCrop as per Udutalapally in combination for a motivation [P.17527-28 Sect.III] “help in the real-time monitoring of the environmental variables… help the edge device layer to perform computations with reduced latency …help the farmer” and further noting [P.17526 ¶3-4] “popularity and advantages of IoT… end goal of smart agriculture.”
However, Udutalapally in combination does not appear to disclose the following limitation which is disclosed by Hadar:
upon receiving input messages via respective feedback connections responsive to the notifications, the input messages comprising images, the {Hadar [0129] “automatically create a geo-tag for indexing the street-level image” again [0128] and [0103,06] “metadata (e.g., a geo-tag, a search index) is automatically created” upon Fig 4 described [0074-76] “remote server and/or computing cloud… connect via network 414” similar at [0157-58]};
Hadar is directed to classification of images with trained deep learning detector models in a distributed cloud system thus being analogous. A person having ordinary skill in the art would have considered it obvious prior to the effective filing date to automatically geo-tag per Hadar in combination for a motivation [0049,52] “automatically creating metadata for indexing” of images so as to “automatically identify …not necessarily be easily determined by a user” and/or such that [0155-56] “discrepancy is detected between the geographic location… When a discrepancy is detected, an automatic correction may be made.”
Hadar further suggests [0099] “correlation between extracted visual features” e.g. [0043] “matching the image(s)” but does not prima facie disclose the following limitations which are disclosed by Harikumar:
correlating the images received via the respective feedback connections in association with the at least first campaign with the at least first type of subject and tagged source metadata comprising the at least one of the one or more defined classifying conditions {Harikumar Fig 4:402 “Correlated images” regards object detection model, Fig 1 network between server and client devices provide respective feedback connections where 106 extraction system extracts the correlated images Fig 2:204, 10:1004 and correlating includes [0067] “extraction system 106 utilizes digital images having matched metadata tags as a series of correlated images. For instance… location metadata tags, and/or user ID metadata tags as the series of correlated images” e.g. [0018] “location metadata within a threshold distance” and classifying conditions include [0099] “classification machine learning model to generated predicted instance labels indicating color and/or content” Figs 2:206, 5:504, similar [0093]}; and
training the at least first model at the central computing network by at least correlating the images received via the respective feedback connections, as components of a first data set for the at least first model, with the at least first type of subject and tagged metadata {Harikumar [0076] “system 106 trains the convolutional neural network 512 to generate image-level labels for the series of correlated images” similar [0070] “detecting objects in the series of correlated images… object detection machine learning model trained on classes from training datasets” e.g. dogs in correlated images to build model Fig 2, training by system 106 shown Fig 1 in communication with network, arrows indicate connection described [0131,35], and tagged metadata as already noted [0067,18]}.
Harikumar is directed to training object detection classification models with a distributed cloud environment thus being analogous. A person having ordinary skill in the art would have considered it obvious prior to the effective filing date to correlate and train per Harikumar in combination to arrive at the invention as claimed for a motivation to [0029] “improve the efficiency and speed for generating similar instance image datasets” which may [0038] “automatically extracts similar instances of objects from a repository of digital images.”
With respect to claim 10, Capota teaches:
A system of expediting distributed feedback for training of supervised learning models to identify one or more types of subjects within images, {Capota Figs 1-3 illustrate system of distributed environment for model training [0024] “Supervised machine learning… for example, classification” to [0130,134] “train an image recognition model… object recognition”} the system comprising:
a central computing network configured, in association with at least a first campaign for development of at least a first model for classifying at least a first type of subject under at least one of one or more defined classifying conditions, {Capota Figs 1-3, 7-8 show network 127, campaign server 125 and devices 122, for [0039] “developing, tuning/optimizing, troubleshooting and managing models” the models to classify e.g. [0134-35] “Classification examples include object recognition (traffic signs, objects in from of a vehicle, etc.), face recognition” cont’d “CNN… image classification” again at [0083,28]. Conditions are disclosed [0131] “light conditions… model needs to be able to recognize, for example, a traffic sign in rainy conditions with the same accuracy as when the sun is shining” and/or [0113,096] “weather conditions”} to
Capota further discloses [0101,103] “push notifications from the campaign server… user defines the campaign through a label and any additional metadata”.
However, Capota does not appear to fairly disclose the image between device and network which is taught by Udutalapally:
generate notifications to each of a plurality of potential feedback source devices requesting responsive images to be received at the central computing network and comprising the least first type of subject {Udutalapally [P.17528 Sect. V.A] “notify famers about the infected crop” illustratively at Fig 3-right “Real-time crop disease prediction notifications” bidirectional arrows note crop image for sCrop mobile application on smartphone devices Fig 13 and in communication with cloud Fig 2, the subject including crop/leaf/plant Figs 4-7, Alg.1},
upon generating the notifications, to establish a respective feedback connection between each of the plurality of potential feedback source devices and a data storage network associated with the at least a first model {Udutalapally [P.1728 ¶1] “5G, WiFi” whereby [P.1727 Last¶] “sCrop device communicates with the upper layers in wireless” describes communication between the devices, shown Figs 2-3 and implemented Alg.1 [P.17531] “SendCloud…ReadCloudtoMobileApp” connects for feeding images read from camera e.g. [P.17529 ¶2] “feed in images of the crops to the Deep Learning (DL) model” and further shows data storage in cloud network at Fig 3},
Udutalapally is directed to training deep learning classifier models with distributed cloud environment thus being analogous. A person having ordinary skill in the art would have considered it obvious prior to the effective filing date to generate notifications over cloud using Udutalapally’s sCrop in combination for a motivation [P.17527-28 Sect.III] “help in the real-time monitoring of the environmental variables… help the edge device layer to perform computation with reduced latency …help the farmer” noting [P.17526 ¶3-4] “popularity and advantages of IoT…end goal of smart agriculture.”
However, Udutalapally in combination does not appear to disclose the following limitation which is disclosed by Hadar:
upon receiving input messages via respective feedback connections responsive to the notifications, the input messages comprising images, automatically the {Hadar [0129] “automatically create a geo-tag for indexing the street-level image” again [0128] and [0103,06] “metadata (e.g., a geo-tag, a search index) is automatically created” upon Fig 4 [0074-76] “remote server and/or computing cloud… connect via network 414” similar at [0157-58]}
Hadar is directed to classification of images with trained deep learning detector models in a distributed cloud system thus being analogous. A person having ordinary skill in the art would have considered it obvious prior to the effective filing date to automatically geo-tag per Hadar in combination for a motivation [0049,52] “automatically creating metadata for indexing” of images to “automatically identify …not necessarily be easily determined by a user” and/or [0155-56] “discrepancy is detected between the geographic location… When a discrepancy is detected, an automatic correction may be made.”
Hadar further suggests [0099] “correlation between extracted visual features” e.g. [0043] “matching the image(s)” but does not prima facie disclose the following limitations which are disclosed by Harikumar:
correlate the images received via the respective feedback connections in association with the at least first campaign with the at least first type of subject and tagged source metadata comprising the at least one of the one or more defined classifying conditions {Harikumar Fig 4:402 “Correlated images” regards object detection model, Fig 1 network between server and client devices provide respective feedback connections where 106 extraction system extracts correlated images Fig 2:204, 10:1004 and correlating includes [0067] “extraction system 106 utilizes digital images having matched metadata tags as a series of correlated images. For instance… matching date metadata tags, location metadata tags, and/or user ID metadata tags as the series of correlated images” e.g. [0018] “location metadata within a threshold distance” and classifying conditions include [0099] “classification machine learning model to generated predicted instance labels indicating color and/or content” Figs 2:206, 5:504, similar [0093]}, and
train the at least first model at the central computing network by least correlating the images received at the central computing network via the respective feedback connections, as components of a first data set for the at least first model, with the at least first type of subject and tagged metadata {Harikumar [0076] “system 106 trains the convolutional neural network 512 to generate image-level labels for the series of correlated images” similarly at [0070] “detecting objects in the series of correlated images… object detection machine learning model trained on classes from training datasets” e.g. dogs in correlated images to build model Fig 2, training by system 106 shown Fig 1 in communication with network, arrows indicate connection described [0131,35], and tagged metadata as already noted [0067,18]}.
Harikumar is directed to training object detection classification models with a distributed cloud environment thus being analogous. A person having ordinary skill in the art would have considered it obvious prior to the effective filing date to correlate and train per Harikumar in combination to arrive at the invention as claimed for a motivation to [0029] “improve the efficiency and speed for generating similar instance image datasets” which may [0038] “automatically extracts similar instances of objects from a repository of digital images.”
With respect to claim 21, Capota teaches:
A method of expediting distributed feedback for training of supervised learning models to identify one or more types of subjects within images {Capota [0024] “techniques may be used for supervised learning, for example, classification” Figs 2-3 and 8 illustrate distributed environment for model management and training [0130,134] “train an image recognition model… object recognition”}, the method comprising:
in association with at least a first campaign for development of at least a first model for classifying at least a first type of subject under at least one of one or more defined classifying conditions {Capota Figs 7-8 show Campaign associated with model repository 135, the campaign being for model development [0039] “developing, tuning/optimizing, troubleshooting and managing models” models to classify e.g. [0134-35] “Classification examples include object recognition (traffic signs, objects in from of a vehicle, etc.), face recognition” cont’d “CNN… image classification” again at [0083,28]. Conditions are disclosed [0131] “light conditions… model needs to be able to recognize, for example, a traffic sign in rainy conditions with the same accuracy as when the sun is shining” and/or [0113,096] “weather conditions” },
Capota further discloses [0101,103] “push notifications from the campaign server… user defines the campaign through a label and any additional metadata”.
However, Capota does not appear to fairly disclose the image between device and network which is taught by Udutalapally:
broadcasting notifications from a central computing network to a plurality of potential feedback source devices requesting responsive images comprising the at least first type of subject of the one or more types of subjects, and under the at least one of the one or more defined classifying conditions {Udutalapally [P.17528 Sect. V.A] “notify famers about the infected crop” shown Fig 3-right “Real-time crop disease prediction notifications” broadcasting is to Fig 2 farmer’s sCrop phone App arrows indicating from cloud and/or edge [P.17530 Last¶] “trained using Google Cloud Services” with cloud send/read implemented Alg.1 [P.17531], using smartphone devices shown Fig 13. The campaign to develop a model is where model is developed by training [P.17529 Sect. VI.B] “training the image classifier model…fit the parameters for training…fine-tune the parameter of the classifier being trained” and which [P.17532 Sect. VII.A] “enables the farmer to choose the crop… selected crop details are sent to the cloud.” Furthermore, classifying conditions may comprise color spectrum Fig 5 and/or [Abst] “weather conditions” The subject including crop/leaf/plant Figs 4-7, Alg.1};
upon generating the notifications, establishing a respective feedback connection between each of the plurality of potential feedback source devices and a data storage network associated with the at least a first model {Udutalapally [P.1717-28 Pg.Brk] “5G, WiFi” as “sCrop device communicates with the upper layers in wireless” describes communication between the devices, shown Figs 2-3 and implemented Alg.1 [P.17531] “SendCloud…ReadCloudtoMobileApp” connects for feeding images read from camera e.g. [P.17529 ¶2] “feed in images of the crops to the Deep Learning (DL) model” and further shows data storage in cloud network at Fig 3};
Udutalapally is directed to training deep learning classifier models with distributed cloud environment thus being analogous. A person having ordinary skill in the art would have considered it obvious prior to the effective filing date to generate notifications over cloud using sCrop as per Udutalapally in combination for a motivation [P.17527-28 Sect.III] “help in the real-time monitoring of the environmental variables… help the edge device layer to perform computations with reduced latency …help the farmer” and further noting [P.17526 ¶3-4] “popularity and advantages of IoT… end goal of smart agriculture.”
However, Udutalapally does not appear to disclose the following limitation which is disclosed by Hadar:
upon receiving input messages via respective feedback connections responsive to the notifications, the input messages comprising images, automatically tagging the images with source-identifying metadata {Hadar [0129] “automatically create a geo-tag for indexing the street-level image” again [0128] and [0103,06] “metadata (e.g., a geo-tag, a search index) is automatically created” upon Fig 4 [0074-76] “remote server and/or computing cloud… connect via network 414” similar at [0157-58]};
Hadar is directed to classification of images with trained deep learning detector models in a distributed cloud system thus being analogous. A person having ordinary skill in the art would have considered it obvious prior to the effective filing date to automatically geo-tag per Hadar in combination for a motivation [0049,52] “automatically creating metadata for indexing” of images to “automatically identify …not necessarily be easily determined by a user” and/or [0155-56] “discrepancy is detected between the geographic location… When a discrepancy is detected, an automatic correction may be made.”
Hadar further suggests [0099] “correlation between extracted visual features” e.g. [0043] “matching the image(s)” but does not prima facie disclose the following limitations which are disclosed by Harikumar:
correlating the images received via the respective feedback connections in association with the at least first campaign with the at least first type of subject and tagged source metadata comprising the at least one of the one or more defined classifying conditions {Harikumar Fig 4:402 “Correlated images” regards object detection model, Fig 1 network between server and client devices provide respective feedback connections where 106 extraction system extracts correlated images Fig 2:204, 10:1004 and correlating includes [0067] “extraction system 106 utilizes digital images having matched metadata tags as a series of correlated images. For instance… matching date metadata tags, location metadata tags, and/or user ID metadata tags as the series of correlated images” e.g. [0018] “location metadata within a threshold distance” and classifying conditions include [0099] “classification machine learning model to generated predicted instance labels indicating color and/or content” Figs 2:206, 5:504, similar [0093] }; and
training the at least first model at the central computing network by at least correlating the images received via the respective feedback connections, as components of a first data set for the at least first model, with the at least first type of subject and tagged metadata {Harikumar [0076] “system 106 trains the convolutional neural network 512 to generate image-level labels for the series of correlated images” similarly at [0070] “detecting objects in the series of correlated images… object detection machine learning model trained on classes from training datasets” e.g. dogs in correlated images to build model Fig 2, training by system 106 shown Fig 1 in communication with network, arrows indicate connection described [0131,35], and tagged metadata as already noted [0067,18]}.
Harikumar is directed to training object detection classification models with a distributed cloud environment thus being analogous. A person having ordinary skill in the art would have considered it obvious prior to the effective filing date to correlate and train per Harikumar in combination to arrive at the invention as claimed for a motivation to [0029] “improve the efficiency and speed for generating similar instance image datasets” which may [0038] “automatically extracts similar instances of objects from a repository of digital images.”
Claims 5 and 22 are rejected under 35 U.S.C. 103 as being unpatentable over Capota, Udutalapally, Hadar and Harikumar in view of Lasso et al., “Discovering weather periods and crop properties for coffee rust incidence from feature selection approaches” hereinafter Lasso.
With respect to claim 5, the combination of Capota, Udutalapally, Hadar and Harikumar teaches the method of claim 1. Lasso teaches wherein
the lighting and/or weather conditions associated with each of the network of available feedback source devices are identified by reference to respective position data with respect to a time of day and/or determined weather conditions, wherein the respective weather conditions for each of the network of available feedback source devices are further determined via one or more third party weather applications and/or databases {Lasso Fig 2 Weather Data from data source i.e. third-party is [P.4 ¶2,4] “CATIE Meteorological Station …From the meteorological station data, we took the following weather variables: …air temperature calculated over the day… relative humidity, number of days” over [P.3 ¶5] “weather time periods, called windows” and gives location as latitude/longitude [P.4 ¶2]. Again Fig 2 shows weather data to generate windows for feature selection based on correlation and models such as random forest. See [P.7 Last2] “sequence of weather conditions” and lighting conditions being [P.4 Last¶] “Shade condition… shaded crop (1) or full sun (0)”}.
Lasso is directed to training classification models with correlation-based feature selection thus being analogous. A person having ordinary skill in the art would have considered it obvious prior to the effective filing date to include weather condition data per Lasso in combination to arrive at the invention as claimed for a motivation [P.3 ¶5] “The objective of this work was to model the CLRI in order to study the impact of weather variables, characterized in different weather time periods” so as for “improving the CLRI modeling… allowed us a more detailed understanding of the favorable conditions for the disease development.”
With respect to claim 22, the combination of Capota,, Udutalapally, Hadar and Harikumar teaches the method of claim 21, and further combination with Lasso teaches the limitation of claim 5. Therefore, the rejection of claim 5 with equal motivation is applied to claim 22.
Claims 7, 20 and 23 are rejected under 35 U.S.C. 103 as being unpatentable over Capota, Udutalapally, Hadar and Harikumar in view of Bhate et al., US PG Pub No 2023/0376805A1 hereinafter Bhate, and further in view of Bourdev et al., US PG Pub No 2015/0036919A1 hereinafter Bourdev.
With respect to claim 7, the combination of Capota, Udutalapally, Hadar and Harikumar teaches the method of claim 1. Bhate teaches comprising
automatically verifying images received at the central computing network via the feedback connection as relating to the at least first type of subject, based at least in part on the source metadata and the corresponding notification {Bhate Fig 7:711 “Auto Validation” i.e. [0044-45] “verification and validation” for model which includes images that can be extracted as a feature vector [0080,73]. Fig 9:939 shows validation set received in a distributed environment, meta-data is shown Fig 4:413 and notification is alert [0078,80]}, and
Bhate is directed to distributed training of classifier models thus being analogous. A person having ordinary skill in the art would have considered it obvious prior to the effective filing date to automatically validate/verify per Bhate in combination for the motivation “performance can be tracked in real time and contemplated against expected values” [0039] and vetting performance can result in rollback or reversion [0107].
However, the following limitation is taught by Bourdev:
only correlating verified images in the data storage, as components of the first data set for the at least first model, with the at least first type of subject and the tagged metadata {Bourdev [0074] “only the top 200 scoring images to the classifier 208 as the training set of images” top-k ranking is described for the correlation module 308, such that verification can be a ranked score of top images, tagged metadata is disclosed [0029]}.
Bourdev is directed to training classifier models thus being analogous. A person having ordinary skill in the art would have considered it obvious prior to the effective filing date to only correlate image of top-k rank per Bourdev in combination to arrive at the invention as claimed for a motivation of “selecting a top scoring subset of the sample set… top scoring subset of the evaluation set” [0088-89].
With respect to claim 20, the combination of Capota, Udutalapally, Hadar and Harikumar teaches the system of claim 10, and further combination with Bhate and Bourdev teaches the limitation of claim 7. Therefore, the rejection of claim 7 with equal motivation is applied to claim 20.
With respect to claim 23, the combination of Capota, Udutalapally, Hadar and Harikumar teaches the method of claim 21, and further combination with Bhate and Bourdev teaches the limitation of claim 7. Therefore, the rejection of claim 7 with equal motivation is applied to claim 23.
Claims 8-9, 11-14, 16-17 and 24 are rejected under 35 U.S.C. 103 as being unpatentable over Capota, Udutalapally, Hadar and Harikumar in view of Heikkila et al., “Unmanned Agricultural Tractors in Private Mobile Networks” hereinafter Heikkila.
With respect to claim 8, the combination of Capota, Udutalapally, Hadar and Harikumar teaches the method of claim 1. Heikkila teaches wherein
the plurality of potential feedback source devices each comprise an onboard display unit, a user interface tool associated with the display unit or with a joystick for controlling one or more vehicle operations, and one or more image sensors, and wherein user engagement of at least one specified portion of the user interface tool responsive to the generated notification causes an image to be captured by at least one of the one or more image sensors and then transmitted via the respective feedback connection {Heikkila [P.8-9] Figs 3-4 shows cockpit of tractor remotely controlled via network at a remote cab [P.8-9 Sect3.2] “the remote control cabin had a steering wheel, pedals, and other controls needed for operating the tractor (Figure 3). Three screens were attached to display video feeds from the tractor” and connectivity with Wifi/5G internet protocols Fig 5 [P.10 ¶1] “Raspberry Pi on the tractor side works as a client to establish a connection to the server Raspberry Pi computer on the remote cab. The idea is that the remote control cabins server has a known address on the internet”}.
Heikkila is directed to distributed environments for image processing thus being analogous. A person having ordinary skill in the art would have considered it obvious prior to the effective filing date to employ interface tools per Heikkila in combination to arrive at the invention as claimed for a motivation being [P.2 ¶2,1] “Our study aims to enable remote control of the tractor interconnecting the in-vehicle CAN of the tractor and the CAN bus of the remote control system” such that the “main idea of the experiment reported here was to establish an ad hoc private network or a private network bubble to enable remote operations of a tractor” e.g. [P.5 Last¶] “smart farms (e.g., demand of being easy to use.”
With respect to claim 9, the combination of Capota, Udutalapally, Hadar, Harikumar and Heikkila teaches the method of claim 8, wherein
the tagged source-identifying metadata for each image comprises a type of image sensor capturing the image, and/or a location of the feedback source devices, and/or a data and time at which the image is captured {Harikumar [0067] “location metadata tags, and/or user ID metadata tags” is source-identifying, date/time is “images having matching date metadata tags… time metadata tags” similar at [0018], [0118]}.
With respect to claim 11, the combination of Capota, Udutalapally, Hadar and Harikumar teaches the system of claim 10, and further combination with Heikkila teaches the limitation of claim 8. Therefore, the rejection of claim 8 with equal motivation is applied to claim 11.
With respect to claim 12, the combination of Capota, Udutalapally, Hadar, Harikumar and Heikkila teaches the system of claim 11, wherein
a subset of the plurality of potential feedback source devices comprises one or more work vehicles {Capota [0008] “subset of devices” Fig 1:122 vehicles, described [0042] “worker devices 122” similar at [0077-78] considers “business” and “worker devices 122 maintained by third-parties”}.
With respect to claim 13, the combination of Capota, Udutalapally, Hadar, Harikumar and Heikkila teaches the system of claim 11, wherein
a subset of the plurality of potential feedback source devices comprises one or more mobile computing devices {Capota [0008] “subset of devices” comprises [0022] “edge/IoT devices” e.g. [0119] “mobile device 122” Fig 2:122}.
With respect to claim 14, the combination of Capota, Udutalapally, Hadar, Harikumar and Heikkila teaches the system of claim 11, and further teaches the limitation of claim 9. Therefore, the rejection of claim 9 is applied to claim 14.
With respect to claim 16, the combination of Capota, Udutalapally, Hadar, Harikumar and Heikkila teaches the system of claim 11, wherein
the computing network is configured to broadcast notifications to a network of available feedback source devices comprising each of the plurality of potential feedback source devices, in accordance with the first campaign, wherein the broadcast notifications include a request for images comprising the at least first type of subject under the at least one of the one or more defined classifying conditions {Udutalapally [P.17528 Sect. V.A] “notify famers about the infected crop” shown Fig 3-right “Real-time crop disease prediction notifications” broadcasting is to Fig 2 farmer’s sCrop App, arrows indicating from cloud and/or edge [P.17530 Last¶] “trained using Google Cloud Services” with cloud send/read implemented Alg.1 [P.17531], using smartphone devices shown Fig 13. The campaign to develop a model is where model is developed by training [P.17529 Sect. VI.B] “training the image classifier model…fit the parameters for training…fine-tune the parameter of the classifier being trained” and which [P.17532 Sect. VII.A] “enables the farmer to choose the crop… selected crop details are sent to the cloud.” Furthermore, classifying conditions may comprise color spectrum Fig 5 and/or [Abst] “weather conditions” The subject including crop/leaf/plant Figs 4-7, Alg.1 }.
With respect to claim 17, the combination of Capota, Udutalapally, Hadar, Harikumar and Heikkila teaches the system of claim 11, wherein
the computing network is configured to select each of the plurality of potential feedback source devices from a network of available feedback source devices based on identified lighting and/or weather conditions associated with each of the network of available feedback source devices as corresponding to the at least one of the one or more defined classifying conditions {Capota shows Fig 6:A120 “Select participating devices” based on campaign A110 where campaign identifies conditions based on classification modeling e.g. [0134-35] “Classification examples include object recognition (traffic signs, objects in from of a vehicle, etc.), face recognition” cont’d “CNN… image classification” again at [0083,28]. Particularly, classifying conditions include [0131] “light conditions… model needs to be able to recognize, for example, a traffic sign in rainy conditions with the same accuracy as when the sun is shining” and/or [0113,096] “weather conditions” }.
With respect to claim 24, the combination of Capota, Udutalapally, Hadar and Harikumar teaches the method of claim 21, and further combination with Heikkila teaches the limitation of claim 8. Therefore, the rejection of claim 8 with equal motivation is applied to claim 24.
Claim 18 is rejected under 35 U.S.C. 103 as being unpatentable over Capota, Udutalapally, Hadar, Harikumar and Heikkila in view of Lasso.
With respect to claim 18, the combination of Capota, Udutalapally, Hadar, Harikumar and Heikkila teaches the system of claim 17, and further combination with Lasso teaches the limitation of claim 5. Therefore, the rejection of claim 5 with equal motivation is applied to claim 18.
Claims 25-26 are rejected under 35 U.S.C. 103 as being unpatentable over Capota, Udutalapally, Hadar, Harikumar and Heikkila in view of Nickl et al., US PG Pub No 2021/0125089A1 hereinafter Nickl.
With respect to claim 25, the combination of Capota, Udutalapally, Hadar and Harikumar teaches the method of claim 21. Nickl teaches comprising
automatically reviewing an image received from a feedback source device, and generating a confirmation feedback request to a display unit associated with the feedback source device regarding a presence of the at least one subject within the received image, wherein confirmation feedback input responsive to the configuration feedback request is attached to the image as metadata for the training of the model {Nickl [0240] “automated image review is generated by a confirmation… supervision over time as the machine learning systems become more deeply trained” further [0268] “image metadata types can be automatically reviewed to identify multiple occurrences of the same image that can be identified with high confidence as not comprising protected data (e.g., logos” shown Fig 1A:108 validate presence of protected data for subsequent incorporation into machine learning 118, Fig 2 distributed system of server-client devices, and example screenshots Figs 3}.
Nickl is directed to distributed system and methods for training machine learning models thus being analogous. A person having ordinary skill in the art would have considered it obvious prior to the effective filing date to automatically review images per Nickl in combination to arrive at the invention as claimed for a motivation of supervising a model’s training data with consideration of protected/sensitive data [0240,268] e.g. [0199] “Such validation can enhance the automated detection of protection by improving the accuracy of protected information identification” and addressing need for compliance [0025].
With respect to claim 26, the combination of Capota, Udutalapally, Hadar and Harikumar teaches the method of claim 21, comprising
automatically reviewing an image received from a feedback source device, and generating a confirmation feedback request to a display unit associated with the feedback source device regarding a presence of the at least one subject within the received image, wherein confirmation feedback input responsive to the configuration feedback request is attached to the image as metadata for the training of the model {Nickl [0240] “automated image review is generated by a confirmation… supervision over time as the machine learning systems become more deeply trained” further [0268] “image metadata types can be automatically reviewed to identify multiple occurrences of the same image that can be identified with high confidence as not comprising protected data (e.g., logos” shown Fig 1A:108 validate presence of protected data for subsequent incorporation into machine learning 118, Fig 2 distributed system of server-client devices, and example screenshots Figs 3}.
Nickl is directed to distributed system and methods for training machine learning models thus being analogous. A person having ordinary skill in the art would have considered it obvious prior to the effective filing date to automatically review images per Nickl in combination to arrive at the invention as claimed for a motivation of supervising a model’s training data with consideration of protected/sensitive data [0240,268] e.g. [0199] “Such validation can enhance the automated detection of protection by improving the accuracy of protected information identification” and addressing need for compliance [0025].
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Li et al., “GeoImageNet: a multi-source natural feature benchmark dataset for GeoAI and supervised machine learning” see Fig 2 labeled metadata, feature selection
Rude et al., PCT WO2022/246548A1 Croptimistic see Fig 1
Adkisson et al., “Autoencoder-based Anomaly Detection in Smart Farming Ecosystem” arXiv: 2111.00099v1 Tennesse Tech, see Fig 1
Mitra et al., “eCrop: A Novel Framework for Automatic Crop Damage Estimation” Figs 3, 5-7.
Bapatla et al., “sFarm: A Distributed Ledger Based Remote Crop Monitoring System for Smart Farming” see Fig 2, Alg.1
Javault et al., US PG Pub No 2021/0149406A1 FarmWise see Figs 1-2
Redden et al., US PG Pub No 2022/0360758A1 Blue River see Fig 3
Delatree et al., US PG Pub No 2023/0363370A1 BASF Agro see Fig 1
Khinvasara et al., US PG Pub No 2023/0237671A1 Nvidia object detection, bounding boxes and regions of interests
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Chase P Hinckley whose telephone number is (571)272-7935. The examiner can normally be reached M-F 9:00 - 5:00.
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, Miranda M. Huang can be reached at 571-270-7092. 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.
/CHASE P. HINCKLEY/Examiner, Art Unit 2124
Prosecution Insights