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 .
Response to Arguments
Applicant’s arguments filed 09/04/2026 have been considered but they are not persuasive.
On pages 6-7, Applicant argues that,
“Applicant submits that Ross, Kinoshita, and Olson, alone or in combination, do not teach or suggest each and every element recited in independent claims 1 and 8. From Applicant's review, Ross appears to teach audio data generating triggering events if raised voices or high levels of vocal stress are detected [0045]. Further, Ross appears to teach progressively overwriting the oldest data in the manner of a circular buffer [0041]. However, Ross does not teach storing a portion of video responsive to an AI model identifying a trigger and deleting a portion of video responsive to the AI model determining that the portion of the video lacks the trigger.
From Applicant's review, the Kinoshita reference does not cure the deficiencies of Ross. Kinoshita appears to teach a monitoring camera that includes AI, a sound collection unit, and a communication unit that receives a parameter for teaching an event of a detection target [0006]. However, Kinoshita does not teach storing a portion of video responsive to an AI model identifying a trigger and deleting a portion of video responsive to the AI model determining that the portion of the video lacks the trigger.”
In response, Examiner respectfully disagrees and submits that, at least in [0046], Ross clearly teaches a model implemented by a recognition algorithm to detect a trigger in portions of the video. In response to the model identifying the trigger, the portion of the video is stored in a non-volatile memory. Otherwise, in response to the model identifying no trigger, the portion of the video stored in the volatile memory is deleted by being overwritten with new data without recording to the non-volatile memory.
Kinoshita is only relied upon to teach implement a trigger detection model as an artificial intelligence (AI) model.
Thus, when substituting the model of Ross by the AI model of Kinoshita, the claimed limitations are met.
Applicant’s other arguments on pages 7-11 are moot in view of the discussion of Ross and Kinoshita above.
Claim Rejections - 35 USC § 103
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.
Claims 1-2, 5-13, and 16 are rejected under 35 U.S.C. 103 as being unpatentable over Ross et al. (US 2015/0063776 A1 – hereinafter Ross) and Kinoshita et al. (US 2020/0226898 A1).
Regarding claim 1, Ross discloses an apparatus (Figs. 2a-2b) comprising: a memory ([0034]; [0041] – a circular buffer implemented using a volatile memory); a camera ([0041] – at least a camera unit); and a processor coupled to the memory and the camera (Fig. 1; [0033]-[0034] – processor 118 coupled to the volatile memory 120 and the camera, e.g. any of the cameras 102 and 104 shown in Fig. 1), wherein the processor is configured to: receiving a model ([0046] – receiving a model implemented by a recognition algorithm); store the model in the memory ([0046] – storing the model in a memory for execution of recognizing a trigger, e.g. a face); record video via the camera ([0041] – recording video captured by the camera); input the video into the model, wherein the model identifies a trigger ([0046] – input the video into a model implemented by a recognition algorithm to identifies a trigger); store a first portion of the video for a first particular time period in the memory ([0041] – storing a first portion of the video in the circular buffer for a first particular time period, which is determined by the length of the circular buffer and the rate at which encoded data is generated); and store a second portion of the video for a second particular time period responsive to the model identifying the trigger in the second portion of the video ([0042]-[0043]; [0048] – responsive to a triggering signal identified in the portion of video by a recognition model as described at least in [0046], storing a second portion of the video for a second particular time period, which is equal to the first portion plus the portion following the event to a non-volatile memory 122); and delete the first portion of the video responsive to the model determining that the first portion of the video lacks the trigger ([0041]-[0042] – in response to no trigger being detected in the video portion by the recognition model as described at least in [0046], the video images are stored in a volatile memory to be deleted by being overwritten with new data).
However, Ross does not disclose the model as an artificial intelligence (AI) model; and training the artificial intelligence (AI) model with a trigger remotely in a cloud device.
Kinoshita discloses a model as an artificial intelligence (AI) model ([0198] – an AI model to detect a trigger in portions of a video); a processor is configured to receive the artificial intelligence (AI) model trained remotely in a cloud device ([0157]-[0158]; [0170] – a processor of a camera receiving an AI model trained remotely in a server device as further described in at least [0153]-[0154], which corresponds to a cloud device); store the AI model in a memory ([0198] – storing the AI model in a storage unit of the camera); record video via a camera (Fig. 11; [0087]; [0133] – recording audio and images via the camera); input the video into the Al model ([0134] – inputting the audio and images into the AI model), wherein the Al model identifies a trigger ([0135]-[0144]; [0290] – the AI model identifies an event, e.g. a harmful animal, a person, a voice of a target etc., that triggers an alarm).
One of ordinary skill in the art before the effective filing date of the claimed invention would have been motivated to incorporate the AI model taught by Kinoshita into the model in the apparatus taught by Ross because it would have been advantageous in updating the model to detect various types of triggering events (Kinoshita: [0213]).
Regarding claim 2, Ross in view of Kinoshita also discloses the apparatus of claim 1, wherein the apparatus is a wearable device ([0006] – the apparatus is worn on the user’s body).
Regarding claim 5, Ross in view of Kinoshita also discloses the apparatus of claim 1, wherein the second particular time period is longer than the first particular time period ([0048] – the second particular time period, which is equal to the first portion plus the portion following the event).
Regarding claim 6, Ross in view of Kinoshita also discloses the apparatus of claim 1, further comprising a global position system (GPS) configured to determine a location of the apparatus ([0035] – comprising sensors include a global-positioning system (GPS) receiver for recording position data of the apparatus).
Regarding claim 7, Ross in view of Kinoshita also discloses the apparatus of claim 6, wherein the memory is configured to store the location of the apparatus responsive to the trigger ([0041]; [0048] – responsive to the trigger, recording the video data combined with the sensor data including the GPS data as described in [0035]).
Regarding claim 8, Ross discloses an apparatus comprising: a camera ([0041] – at least a camera unit); a memory ([0034]; [0041] – a memory comprising a volatile memory 120 and a non-volatile memory 122); and a processor coupled to the memory and the camera (Fig. 1; [0033]-[0034] – processor 118 coupled to the volatile memory 120, the non-volatile memory 122, and the camera, e.g. any of the cameras 102 and 104 shown in Fig. 1), wherein the processor is configured to: receiving a model ([0046] – receiving a model implemented by a recognition algorithm); store the model in the memory ([0046] – storing the model in a memory for execution of recognizing a trigger, e.g. a face); record video via the camera ([0041] – recording video captured by the camera); record video via the camera ([0041] – recording video captured by the camera); input the video into a model, wherein the model identifies whether portions of the video include a trigger ([0046] – the processor executing a model implemented by a recognition model to detect a trigger from the video inputted to the model); store a first portion the video in the memory responsive to the model identifying the trigger in first second portion of the video ([0042]-[0043]; [0048] – responsive to a triggering signal identified in a portion of video by the recognition model as further described at least in [0046], storing a first portion of the video to a non-volatile memory 122); and delete a second portion of the video responsive to the model determining that the second portion of the video lacks the trigger ([0041]-[0042] – in response to no trigger being detected in a portion of video by the recognition model as further described at least in [0046], the video images are stored in a volatile memory to be deleted by being overwritten with new data).
However, Ross does not disclose the model as an artificial intelligence (AI) model; the processor is configured to receive the artificial intelligence (AI) model trained remotely in a cloud device; store the AI model in the memory; input the video into the Al model, wherein the Al model identifies whether portions of the video include a trigger.
Kinoshita discloses a model as an artificial intelligence (AI) model ([0198] – an AI model to detect a trigger in portions of a video); a processor is configured to receive the artificial intelligence (AI) model trained remotely in a cloud device ([0157]-[0158]; [0170] – a processor of a camera receiving an AI model trained remotely in a server device as further described in at least [0153]-[0154], which corresponds to a cloud device); store the AI model in a memory ([0198] – storing the AI model in a storage unit of the camera); input the video into the Al model ([0134] – inputting the audio and images into the AI model), wherein the Al model identifies whether portions of the video include a trigger ([0135]-[0144]; [0290] – the AI model identifies an event, e.g. a harmful animal, a person, a voice of a target etc., that triggers an alarm).
One of ordinary skill in the art before the effective filing date of the claimed invention would have been motivated to incorporate the AI model taught by Kinoshita into the model in the apparatus taught by Ross because it would have been advantageous in updating the model to detect various types of triggering events (Kinoshita: [0213]).
Regarding claim 9, Ross also discloses the apparatus of claim 8, wherein the memory further comprises volatile memory ([0041] - volatile memory 120).
Regarding claim 10, Ross also discloses the apparatus of claim 9, wherein the processor is configured to store the video in the volatile memory ([0041] – storing the video in the circular buffer implemented on the volatile memory).
Regarding claim 11, Ross also discloses the apparatus of claim 8, wherein the processor is configured to store a different portion of the video responsive to the output of the AI model ([0048] – storing a portion following the event responsive to the output of the model, which is, in view of Kinoshita as discussed above, the AI model).
Regarding claim 12, Ross also discloses the apparatus of claim 11, wherein the memory further comprises non-volatile memory, and wherein the processor is configured to store the different portion of the video in the non-volatile memory ([0048] – storing the portion following the event onto the non-volatile memory 122).
Regarding claim 13, Ross also discloses the apparatus of claim 8, wherein the processor is configured to transmit a different portion of the video responsive to the output of the AI model ([0048] – transmitting the portion following the event onto the non-volatile memory 122 responsive to the output of the model, which is, in view of Kinoshita discussed above, the AI model).
Regarding claim 16, Ross discloses a method comprising: receiving a model at a computing device ([0046] – receiving a model implemented by a recognition algorithm at a computing device); storing the model in memory of the computing device ([0046] – storing the model in memory of the computing device); record video via the camera of the computing device ([0041] – recording video captured by the camera); input the video into a model at the computing device, wherein the model identifies a trigger ([0046] – the processor executing a model implemented by a recognition model to detect a trigger from the video inputted to the model); store a first portion the video in the memory responsive to the model identifying the trigger in first second portion of the video ([0042]-[0043]; [0048] – responsive to a triggering signal identified in a portion of video by the recognition model as further described at least in [0046], storing a first portion of the video to a non-volatile memory 122); and delete a second portion of the video responsive to the model determining that the second portion of the video lacks the trigger ([0041]-[0042] – in response to no trigger being detected in a portion of video by the recognition model as further described at least in [0046], the video images are stored in a volatile memory to be deleted by being overwritten with new data).
However, Ross does not disclose the model as an artificial intelligence (AI) model; training the AI model with a trigger remotely in a cloud device; the processor is configured to receive the artificial intelligence (AI) model trained remotely in a cloud device; store the AI model in the memory; input the video into the Al model, wherein the Al model identifies whether portions of the video include a trigger.
Kinoshita discloses a model as an artificial intelligence (AI) model ([0198] – an AI model to detect a trigger in portions of a video); and training the artificial intelligence (AI) model with a trigger remotely in a cloud device ([0153]-[0154] – training an AI model with a trigger remotely in a server device, which corresponds to a cloud device); storing the AI model at a computing device ([0198] – storing the AI model in a storage unit of a computing device comprising a camera); recording video via a camera of the computing device (Fig. 11; [0087]; [0133] – recording audio and images via the camera); input the video into the Al model at the computing device ([0135] – inputting the audio and images into the AI model at the computing device comprising the camera), wherein the Al model identifies the trigger ([0135]-[0144]; [0290] – the AI model identifies an event, e.g. a harmful animal, a person, a voice of a target etc., that triggers an alarm).
One of ordinary skill in the art before the effective filing date of the claimed invention would have been motivated to incorporate the teachings of Kinoshita into the method taught by Ross because it would have been advantageous in updating the model to detect various types of triggering events (Kinoshita: [0213]).
Claims 3 and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Ross and Kinoshita as applied to claims 1-2, 5-13, and 16 above, and further in view of Wang (CN 110889354 A – hereinafter Wang, references to machine translated copy previously attached).
Regarding claim 3, the see teachings of Ross and Kinoshita as discussed in claim 2 above. However, Ross and Kinoshita do not disclose the wearable device is augmented reality (AR) glasses.
Wang discloses a wearable device is augmented reality (AR) glasses (page 2).
One of ordinary skill in the art before the effective filing date of the claimed invention would have been motivated to incorporate the teachings of Wang into the apparatus taught by Ross and Kinoshita to implement the apparatus in form of AR glasses which were known to have been advantageous due to their ability to assist patrol officers immediately identify suspects on the street, their ability to generate overlaid 3D maps on building floor plans and transport routes to enhance situational awareness, etc.
Regarding claim 17, see the teachings of Ross and Kinoshita as discussed in claim 16 above. However, Ross and Kinoshita do not disclose displaying a different video on a user interface
Wang discloses displaying a different video on a user interface (page 3 - according to automatically capture image playback command, playing active trigger captured unread image by the AR glass user, thus displaying the unread image, which comprises the video as described on page 8).
One of ordinary skill in the art before the effective filing date of the claimed invention would have been motivated to incorporate the teachings of Wang into the method taught by Ross and Kinoshita to implement the apparatus in form of AR glasses which were known to have been advantageous due to their ability to assist patrol officers immediately identify suspects on the street, their ability to generate overlaid 3D maps on building floor plans and transport routes to enhance situational awareness, etc.
Claims 14-15 are rejected under 35 U.S.C. 103 as being unpatentable over Ross and Kinoshita as applied to claims 1-2, 5-13, and 16 above, and further in view of Wiklof (US 2009/0324203 A1 – hereinafter Wiklof).
Regarding claim 14, see the teachings of Ross and Kinoshita as discussed in claim 13 above. However, Ross and Kinoshita do not disclose the processor is configured to transmit the different portion of the video to an external device.
Wiklof discloses a processor is configured to transmit a different portion of video to an external device ([0052]-[0053] – transmitting a portion of video to a remote storage device).
One of ordinary skill in the art before the effective filing date of the claimed invention would have been motivated to incorporate the teachings of Wiklof into the apparatus taught by Ross and Kinoshita to enhance storage scalability, i.e. increasing or decreasing the storage capacity without affecting the hardware configuration.
Regarding claim 15, see the teachings of Ross, Kinoshita, and Wiklof as discussed in claim 14 above, in which Wiklof also discloses the external device is a cloud device or computing device ([0042]; Fig. 4 – a server or a client computer).
The motivation for incorporating the teachings of Wiklof into the apparatus has been discussed in claim 14 above.
Claim 18 is rejected under 35 U.S.C. 103 as being unpatentable over Ross, Kinoshita, and Wang as applied to claims 1-3, 5-13, and 16-17 above, and further in view of Moon (US 2017/0289224 A1 – hereinafter Moon).
Regarding claim 18, see the teachings of Ross, Kinoshita, and Wang as discussed in claim 17 above. However, Ross, Kinoshita, and Wang do not disclose providing a selection to store or delete the different video to a user responsive to displaying the different video on the user interface.
Moon discloses providing a selection to store or delete a video to a user responsive to displaying the video on a user interface ([0034]).
One of ordinary skill in the art before the effective filing date of the claimed invention would have been motivated to incorporate the teachings of Moon into the different video in the method taught by Ross, Kinoshita, and Wang to allow the user to correct any error made by the AI model by deleting unwanted videos.
Claims 19-20 are rejected under 35 U.S.C. 103 as being unpatentable over Ross, Kinoshita, Wang, and Moon as applied to claims 1-3, 5-13, and 16-18 above, and further in view of Ashlock et al. (US 2021/0350011 A1 – hereinafter Ashlock).
Regarding claim 19, see the teachings of Ross, Kinoshita, Wang, and Moon as discussed in claim 18 above. However, Ross, Kinoshita, Wang, and Moon do not disclose identifying characteristics of the different video.
Ashlock discloses identifying characteristics of a content ([0047] – identifying features of content that is not accepted).
One of ordinary skill in the art before the effective filing date of the claimed invention would have been motivated to incorporate the teachings of Ashlock into the different portion of video discussed in claim 18 above in order to re-train the AI model by providing the AI model with features identified as unwanted by the user, thus enhancing the accuracy of the model.
Regarding claim 20, see the teachings of Ross, Kinoshita, Wang, Moon, and Ashlock as discussed in claim 19 above, in which Ashlock in view of Wang also discloses training the AI model using the selection and the characteristics of the different video ([0047] - re-training a machine learning model with updated training sets to weaken associations between characteristics of the parent entity and the recommended feature responsive to the parent entity not accepting the recommendation).
The motivation for incorporating the teachings of Ashlock has been discussed in claim 19 above.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to HUNG Q DANG whose telephone number is (571)270-1116. The examiner can normally be reached IFT.
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, Thai Q Tran can be reached at 571-272-7382. 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.
/HUNG Q DANG/Primary Examiner, Art Unit 2484
