DETAILED ACTION
Claim Status
Claims 1, 8, and 15 have been amended. Thus, Claims 1, 3-8, and 10-21 are currently pending.
Response to Amendments
Applicant’s remarks and amendments filed October 30, 2025, have been entered.
Response to Arguments
Applicant’s arguments filed October 30, 2025, regarding the rejection(s) of claim(s) 1, 3-8, and 10-21 have been fully and completely considered but are moot because the arguments do not apply to the new combination of the references, facilitated by Applicant’s newly submitted amendments, including new prior art— Wu et al, US 20150206014 —being used in the current rejection.
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, 3-8, and 10-20 are rejected under 35 U.S.C. 103 as being unpatentable over Ng et al (“Lightweight deep neural network approach for parking violation detection”) in view of Wu et al, US 20150206014 in view of Yan et al, WO2021203717A1 in further in view of Ratti et al, US 11037024.
With respect to Claims 1 and 8: Ng teaches
A method for automated detection of parking violations, comprising: [Ng has disclosed (Fig 1 page 333, abstract, page 333 RHC section 3 Final paragraphs) has disclosed an image processing system have at least a camera device for acquiring the image data and a base station for processing the image data. The device of Ng (Raspberry-Pi device – see abstract of NG, said Raspberry-Pi type device of Ng inherently comprising processor, memory, and software for performing the disclosed processes) inherently comprising processor, memory, and software for performing the disclosed processes.]
capturing image data using an edge device equipped with a camera, wherein the image data includes a plurality of frames of a video; [Ng (page 332 RHC final paragraph) has disclosed an edge computing device having an IP camera for acquiring video image data processed by the system and method of Ng.]
identifying, using a first semi-supervised learning model [Ng (page 334 LHC 1st full paragraph) has disclosed a Self-Prepared “semi-supervised” learned classification model using MobileNet, wherein the sample training regions of the image data are manually extracted into labeled classes (vehicle and background type labeled classes of the image regions). self-training is a semi-supervised learning technique ], at least one vehicle in the image data and a position of the at least one vehicle in the image data [Ng (page 334 LHC 1st and 2nd full paragraphs) has disclosed extracting, using MobileNet, vehicle regions from the monitored IPR regions of the captured image data.]
[Ng has disclosed locating vehicles in regions of the captured video image data, but has not further disclosed a second semi-supervised learning model for identifying rear vehicle light(s) on the identified vehicle as required by the below presented claim limitation.]
identifying, using a [Ng, is relied upon for supplying the “semi-supervised” training samples that are labeled “semi-supervised” by a user for training of the models (see above discussion of Ng).]
and determining whether the at least one vehicle is in a waiting state by detecting at least one condition within a window of [Ng [pg. 334, 3. System Design] To suppress false alarms due to vehicles passing by or entering and leaving parking spaces, temporal hysteresis technique is incorporated to make sure a stationary vehicle is reported only when it is detected at the same position for at least three consecutive cycles. A window of at least three consecutive cycles is a time frame of where parking lot images are periodically received]
determining a double parking status of the at least one vehicle based on the position of the at least one vehicle [Ng (page 333 RHC final paragraph through 1st paragraph LHC page 334) has disclosed determining and displaying parking violations based on the parking region and vehicle data of said parking regions, wherein Ng further utilizes vehicle position data to determine parking violations including double parking violations.]
Ng fails to teach
and determining if the at least one vehicle is a double parking candidate based on whether the at least one vehicle is in a road next to a legally parked vehicle in the image data;
when the at least one vehicle is determined to be a double parking candidate,
and determining whether the at least one vehicle is in a waiting state by detecting at least one condition within a window of five frames of the image data,
the at least one condition including a flash of the at least one rear vehicle light
monitoring a status of the at least one rear vehicle light over the plurality of frames;
and determining a double parking status of the at least one vehicle based on the position of the at least one vehicle and the status of the at least one rear vehicle light over the plurality of frames by determining whether the at least one vehicle is in the waiting state indicating that the at least one vehicle is stationary and not double parked when the flashing of hazard lights is detected,
However, Wu teaches
and determining if the at least one vehicle is a double parking candidate based on whether the at least one vehicle is in a road next to a legally parked vehicle in the image data; [see Wu, Fig. 1B, Fig.5, and Paragraph [0029], “In the illustrated example, the enforcement area is a driving lane located next to the parking lane. Where the parking area may be full along one extent and there are no available spaces, vehicles may be tempted to temporarily park and/or stop in the driving lane near or next to the parked vehicles,” parking area contains legally parked vehicles, enforcement area is where vehicle is double parked next to legally parked vehicles]
when the at least one vehicle is determined to be a double parking candidate, [see Wu, Fig. 2, S14, confirming candidate vehicle is double parked using evidence of event and whether candidate vehicle qualifies for an exception or not]
and determining whether the at least one vehicle is in a waiting state by detecting at least one condition within a window of five frames of the image data, [see Wu, Paragraph [0031], “The amount of time that the detected vehicle remains stationary is estimated by counting a number of frames the detected vehicle does not move at S10,” counting a number of frames is considered to include a window of five frames of image data]
the at least one condition including a flash of the at least one rear vehicle light, as a flashing of hazard lights of the least one vehicle [see Wu, Paragraph [0064], “This event can include the operation of hazard (warning) lights (flashers) on the detected vehicle or another detected vehicle behind the detected vehicle, which indicates a hazard such as the vehicle being stopped in or near moving traffic,” and Paragraph [0068], “The module 122 identifies a hazard light region in the sequence of frames surrounding one of a front light and rear light area on the detected vehicle at S622.”]
monitoring a status of the at least one rear vehicle light over the plurality of frames; [see Wu, Paragraph [0068], “The module 122 identifies a hazard light region in the sequence of frames surrounding one of a front light and rear light area on the detected vehicle at S622. A pixel analysis is performed within this region, referred hereafter as a (e.g., lower right) quadrant,” monitoring the status of a hazard light region in the sequence of frames which includes rear light area is considered to be monitoring a status of the at least one rear vehicle light over the plurality of frames]
and determining a double parking status of the at least one vehicle based on the position of the at least one vehicle and the status of the at least one rear vehicle light over the plurality of frames by determining whether the at least one vehicle is in the waiting state indicating that the at least one vehicle is stationary and not double parked when the flashing of hazard lights is detected, [see Wu, Fig. 7A and Paragraph [0070], “However, the hazard lights may indicate that the vehicle requires attention. In one embodiment, in response to a determination that the vehicle is operating its hazard lights, the module 122 determines whether the stationary vehicle is of a type that qualifies for an exception at S630. In response to the stationary vehicle qualifying for an exception (YES at S632), the module 122 classifies the stationary vehicle as not being double parked and not in violation of a traffic rule at S628”)]
[The combination of Ng and Wu are analogous art because they both all in the same field of detecting double parked vehicles. Therefore, it would have been obvious to persons of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the teaching of Ng’s method for automated detection of parking violations with the teaching of Wu’s parking area to determine legally parked vehicles and enforcement area to determine double parked candidates and monitoring a number of frames to determine vehicle status with hazard rear lights. The motivation to combine the teachings of Ng and Wu discloses to reduce costs and improve efficiency, municipalities are exploring the use of new technologies for automating exclusion zone enforcement (see Wu, Paragraph [0003]).]
Ng in view of Wu fails to teach
identifying, using a second semi-supervised learning model, at least one rear vehicle light on the at least one vehicle
However, Yan teaches
identifying, using a second semi-supervised learning model, at least one rear vehicle light on the at least one vehicle [Yan (0034-0035 and 0091-0092) has disclosed identifying a vehicle in a parking space area using a first detection module, then identifying tail/rear/brake type lights of the vehicle in video image data (para 0037-0038 and 0094-0096) using a second detection module, then further performing a determination of parking behavior based on the detection result of the first and second detection modules (para 0038-0039, behavior – para 0108-0110). Yan (at last para 0080-0082) discloses that the rear/brake/tail light type objects of the vehicle are further identified on said already identified vehicle in order to improve the following and tracking of vehicle behavior based on a confirmation of vehicle position using a second vehicle detection process based on a tail light of said vehicle. Yan (para 0084-0088) has further disclosed that the method of each of the first and second detection modules are trained models of received training samples. Thus the combined teachings of Ng and Yan have disclosed this claimed limitations, wherein the motivation and reasons for combining have been disclosed below.]
[Ng, Wu, and Yan are analogous art of image data processing to identify and monitor a vehicle over a sequence of received video image data, based on received identification information of the vehicle being monitored. It would have been obvious to one of ordinary skill in the art at the time of the invention to at least try to modify the vehicle only identification for monitoring of a vehicle in video image data of Ng in view of Wu to further identify rear/brake/tail lights of the identified vehicle to perform monitoring of the vehicle based further on a rear/brake/tail light as disclosed by Yan. The motivation for combining would have been to improve the identification and monitoring of a vehicle in video image data in an outdoor scene as required by the teachings of both references, using an enhanced set of identified features of a tail/rear/brake light type identified feature as disclosed by Yan (reduce misdetections – para 0081 of Yan). Therefore, it would have been obvious to one of ordinary skill in the art at the time of the invention to at least try to combine the teachings of Ng in view of Wu with Yan to current limitations of the presented claim language.]
Ng in view of Wu in view of Yan fails to teach
wherein the edge device is installed on a transportation vehicle configured to traverse an area while performing automated detection and data uploading of parking violations.
However, Ratti teaches
wherein the edge device is installed on a transportation vehicle configured to traverse an area while performing automated detection and data uploading of parking violations. [Ratti (Fig 1, col 5 lines 35-40) an edge based mobile device (col 13 lines 48-60 – cell phone or mobile phone type device and vehicle mounted cameras) for detection of objects and determination of parking violations of imaged objects such as vehicles (col 5 lines 40-50),]
[The combination of Ng, Wu, Yan, and Ratti are analogous art because they are all in the same field of traffic violation detection. Therefore, it would have been obvious to persons of ordinary skill in the art, before the effective filing date of the claimed invention to combine the teaching of Ng in view of Wu in view of Yan method for automated detection of parking violations with the teaching of Ratti’s edge based mobile device. The motivation to combine the teachings of Ng in view of Wu in view of Yan and Ratti discloses to provide a technological solution to a major and common problem of inefficient and illegal parking (see Ratti, Col 11, Lines, 58-62).]
With respect to Claim 3: The method of claim 1, further comprising
training the first semi-supervised learning model and the second semi-supervised learning model using unlabeled image data. [The training of machine learning using unlabeled data that is then at least labeled (Ng – page 334 RHC 1st full paragraph) and cropped to create the image processing library of training images, wherein Ng is relied upon to teach the training sample image data of the learning model of the vehicle identification process.]
With respect to Claim 4: The method of claim 1, wherein
the first semi-supervised learning model includes at least one neural network. [Ng – deep learning neural network (abstract).]
With respect to Claims 5 and 11: The method of claim 1, wherein
the first semi-supervised learning model are quantized. [The image processing on a digital computing system (abstract of Ng) such as disclosed by Ng is at least quantized to a binary format.]
With respect to Claims 6 and 13: The method of claim 1, further comprising
the edge device determining a location of the at least one vehicle and associating the location of the at least one vehicle with the double parking status of the at least one vehicle. [Ng the base station edge type computing device calculates and generates data (Ng page 333 RHC final paragraph through 1st paragraph LHC page 334) for display that is at least parking violation data, location information, and captured image data of the illegally parked vehicle (Fig 3 page 335 of Ng).]
With respect to Claims 7 and 14: The method of claim 1, further comprising
the edge device saving the image data and double parking status of the at least one vehicle. [The edge computing type base station (Ng page 333 RHC final paragraph through 1st paragraph LHC page 334) at least temporarily storing the data for delivery to the mobile phone or could device for access and display (at least Fig 3 page 335 of Ng).]
With respect to Claim 10: The device of claim 8, wherein
the edge device is a mobile device. [Ratti (Fig 1, col 5 lines 35-40) an edge based mobile device (col 13 lines 48-60 – cell phone or mobile phone type device and vehicle mounted cameras) for detection of objects and determination of parking violations of imaged objects such as vehicles (col 5 lines 40-50), wherein the image processing is a machine learning type image processing (AI of Yan – col 5 lines 35-50). The mobile device being mounted and part of an image processing device mounted on a mobile vehicle (col 13 lines 48-60 of Yan).]
With respect to Claim 12: The edge device of claim 8, wherein
the processing circuitry is further configured to modify the image data. [Ng (Ng page 333 RHC final paragraph through 1st paragraph LHC page 334 and Fig 3 page 335) has disclosed image processing to modify the image data to crop to a vehicle region identifying the particular vehicle violating the parking rules.]
As per Claim 15, Claim 15 a system for automated detection of parking violations as claimed in Claim 1. Therefore the rejection and rationale are analogous to that made in Claim 1.
Claim 15 further recites
at least one camera; [Ng – Fig 1]
a transportation vehicle; [Ratti (Fig 1, col 5 lines 35-40) an edge based mobile device (col 13 lines 48-60 – cell phone or mobile phone type device and vehicle mounted cameras) for detection of objects and determination of parking violations of imaged objects such as vehicles (col 5 lines 40-50), wherein the image processing is a machine learning type image processing (AI of Yan – col 5 lines 35-50). The mobile device being mounted and part of an image processing device mounted on a mobile vehicle (col 13 lines 48-60 of Yan).]
at least one device comprising circuity configured to [See above device of Ng, and Ng in view of Wu in view of Yan in view of Ratti: Ratti having further disclosed the mobile processing device that is the edge device mounted on a transportation vehicle.]
receive, from the at least one camera, image data comprising a plurality of frames of a video; [Ng (page 332 RHC final paragraph) has disclosed an edge computing device having an IP camera for acquiring video image data processed by the system and method of Ng.]
With respect to Claim 16: The system of claim 15, wherein
the at least one device is an edge device. [Ng (page 332 RHC final paragraph) has disclosed an edge computing device having an IP camera for acquiring video image data processed by the system and method of Ng.]
With respect to Claim 17: The system of claim 15, wherein
the first semi-supervised learning model and the second semi-supervised model are quantized. [The image processing on a digital computing system (abstract of Ng) such as disclosed by Ng is at least quantized to a binary format.]
With respect to Claim 18: The system of claim 15, wherein
the transportation vehicle is an unmanned vehicle. [Ratti (col 6 lines 10-20) has disclosed that the vehicle can be a drone type vehicle.]
With respect to Claim 19: The system of claim 15, wherein
the processing circuitry is further configured to determine a location of the at least one vehicle and associate the location of the at least one vehicle with the double parking status of the at least one vehicle. [Ng the base station edge type computing device calculates and generates data (Ng page 333 RHC final paragraph through 1st paragraph LHC page 334) for display that is at least parking violation data, location information, and captured image data of the illegally parked vehicle (Fig 3 page 335 of Ng).]
With respect to Claim 20: The system of claim 15, wherein
the processing circuitry is further configured to save the image data and the double parking status of the at least one vehicle. [The edge computing type base station (Ng page 333 RHC final paragraph through 1st paragraph LHC page 334) at least temporarily storing the data for delivery to the mobile phone or could device for access and display (at least Fig 3 page 335 of Ng).]
Claim 21 is rejected under 35 U.S.C. 103 as being unpatentable over Ng et al (“Lightweight deep neural network approach for parking violation detection”) in view of Wu et al, US 20150206014 in view of Yan et al WO2021203717A1 in view of Ratti et al, US 11037024 and further in view of Weng et al. (Learning with not Enough Data Part 1: Semi-Supervised Learning; available at: https://lilianweng.github.io/posts/2021-12-05-semi-supervised/).
With respect to Claim 21: The method of claim 3, wherein
Ng in view of Wu in view of Yan in further in view of Ratti fails to teach
the training includes training the first semi-supervised learning model using a teacher model and the unlabeled image data includes augmented image data
However, Weng teaches
the training includes training the first semi-supervised learning model using a teacher model and the unlabeled image data includes augmented image data. [[pg. 5, Mean teachers] Input augmentation (e.g. random flips of input images, Gaussian noise) or student model dropout is necessary for good performance. Dropout is not needed on the teacher model.]
[Therefore, it would have been obvious to persons of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the teaching of Ng in view of Wu in view of Yan in further in view of Ratti’s method for automated detection of parking violations with the teaching of Weng’s the training includes training the first semi-supervised learning model using a teacher model and the unlabeled image data includes augmented image data. The motivation to combine the teachings of Ng in view of Wu in view of Yan in further in view of Ratti and Weng is because including the teachings of Weng enhances the method by avoiding errors cause by limited labeled data for learning tasks (Weng [pg. 1]).]
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 DOMINIQUE JAMES whose telephone number is (703)756-1655. The examiner can normally be reached 9:00 am - 6:00 pm EST.
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, Emily Terrell can be reached at (571)270-3717. 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.
/D.J./Examiner, Art Unit 2666
/MING Y HON/Primary Examiner, Art Unit 2666
Prosecution Insights