DETAILED ACTIONS
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 Amendment
This office action is in response to the amendments/arguments submitted by the Applicant(s) on 03/23/2026. This is a second non final rejection.
Response to Arguments
Status of the Claims
Claims 1-21 are pending.
Claims 12 and 13 are amended.
Claims 22-42 cancelled.
Rejections Under 35 U.S.C. §103
Applicant's argument/amendment, see remarks pages 12-16, filed 03/23/2026
with respect to the rejection(s) of Claims under 35 U.S.C. §103 have been fully considered and are persuasive. Therefore, the rejection(s) has been withdrawn. However, a new rejection is set forth below with a newly found prior art.
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 1-11, and 14-21 are rejected under 35 U.S.C. 103 as being unpatentable over Yang et al. (US 2020/0401883 A1, hereinafter Yang) and in view of Lee Kamp Redden (US 2020/0187406 A1, hereinafter Redden).
Regarding claim 1, Yang teaches,
A method for detecting one or more plants in a sequence of images, the method performed by a processor (Yang, Figure 1, [0039] Plant recognition engine 116 may be configured to perform various aspects of the present disclosure on vision
data captured from various sources, such as people, robots 108, etc., [0058], predictions may be determined using a time-sequence of images of the same plant, which may be associated with each other (e.g., in database 114) and/or with a unique identifier of the plant using techniques described herein”) Figure 6, [0073] At block 602, the system may obtain a digital image that captures at least a first plant of a plurality of plants. Figure 7, processor 714, [0076], Computing
device 710 typically includes at least one processor 714) and comprising:
detecting one or more plants in a first image of the sequence of images by, for at least a first plant of the one or more plants (Yang, Figure 6, [0039], “plant recognition engine 116 may be configured to obtain a digital image that captures at least a first plant of a plurality of plants”. Figure 3, time sequence images):
generating a first detection region for the first plant based on the first image (Yang, Figure 3-4, [0063],”as depicted in FIG. 3, the latent space embeddings
446 1_2 generated from the anchor image 4421 and positive image 4422-both depicting the same plant 4301-are fairly similar, and therefore should be close
together in latent space”. NOTE 446 is interpreted as detection region for the first plant based on first image);
Yang teaches detecting plant growth by training and applying machine learning models to digital images. Yang is silent on selecting image region and image trackers.
Yang is silent on initializing a first tracker for the first plant based on the first detection region, the first tracker having a first state; and
generating a first tracking region by the first tracker for the first plant based on the first state;
detecting at least the first plant in a second image of the sequence of images by, for at least the first plant:
updating the first tracker to have a second state based on the second image and the first state; and generating a second tracking region based on the second state.
However, Redden teaches initializing a first tracker for the first plant based on the first detection region, the first tracker having a first state (Redden,Figure 2-5 [0019] As shown in FIGS. 2 to 5, identifying individual plants includes identifying a foreground region within the image S110, identifying points of interest within the foreground region S120, classifying the points of interest as plant centers or non-plant centers S130”); and
generating a first tracking region by the first tracker for the first plant based on the first state (Redden, [0019] segmenting the foreground region into sub-regions S140, wherein each sub-region encompasses a single point of interest classified
as a plant center. The image from which the individual plants can be a frame of a video or a static image. Each image can be analyzed independently (diachronically) and/or in comparison to other images to account for changes in time ( e.g.synchronically). [0021] When a point of interest is identified, the position of the point of interest within the image is preferably recorded along with the defining features of the point of interest ( e.g. gradient curvature, saturation,
area, shape, etc.). Points of interest are preferably extracted from every foreground region identified in the image”) ;
detecting at least the first plant in a second image of the sequence of images by, for at least the first plant: updating the first tracker to have a second state based on the second image and the first state; and generating a second tracking region based on the second state. (Redden, [0022] a default confidence level S132 (as shown in FIG. 3A), which can be subsequently adjusted. The confidence level for the point of interest is preferably increased if the point of interest is identified or extracted from a predetermined area or pixel-neighborhood (after accounting for movement of the system) in a subsequent image S134 (as shown in FIG.3B), and preferably decreased otherwise. The confidence level for each point of interest is preferably updated with each new image. [0024] wherein each sub-region preferably encompasses a single point of interest classified as a plant center and can
include any suitable number of points of interest classified as non-plant center)
It would have been obvious to a person of ordinary skill before the effective filing date to modify Yang’s method in view of Redden to include a tracking area associated with the state of each target state in the sequence of images and update the tracking state based on second image state as taught by Redden with the benefit of detecting and tracking a target objects from the one or more targets from the plurality of targets based on a target state associated with each of the one or more target image (redden , [0019]-[0022]).
Regarding claim 2, combination of Yang and Redden teaches the method according to claim 1.
Yang is silent on wherein a center of mass of the plant is substantially stationary between the first and second images.
However, Redden teaches wherein a center of mass of the plant is substantially stationary between the first and second images.(Redden, [0021] the plant center for the foreground regions indicative of a single plant can be determined as the
centroid of the region, a point equidistant from all edges of the region, a point randomly selected within the region, or any other suitable point within the region”).
It would have been obvious to a person of ordinary skill before the effective filing date to modify Yang’s method in view of Redden to include a tracking area associated with the state of each target state in the sequence of images and update the tracking state based on second image state as taught by Redden with the benefit of detecting and tracking a target objects from the one or more targets from the plurality of targets based on a target state associated with each of the one or more target image (Redden , [0019]-[0022]).
Regarding claim 3, combination of Yang and Redden teaches the method according to claim 2.
Yang further teaches wherein the first plant exhibits growth between the first and second images (Yang, Figure 3 shows images of the same plant at 15 days and at 25 days, and shows the growth of the same plant).
Regarding claim 4, combination of Yang and Redden teaches the method according to claim 1.
Yang further teaches wherein detecting at least the first plant in the second image further comprises: generating a second detection region for the first plant based on the second image (Yang, Figure 3, [0060], Ten days later, a similar, if larger, bounding shape 3364 is detected around first plant 2301. Similar to bounding shape 3363 , bounding shape 3364 captures, in addition to the various subtle visual features of first plant 2301 itself, the rich environmental context of first plant 2301, including first sprinkler head 234”NOTE: the boundary shape are detected area with attributes related to the plant in the image) ;
Yang is silent on updating the first tracker to have an updated second state based on the second detection region and the second state; and generating an updated second tracking region based on the second state.
However, Yang and Redden teaches updating the first tracker to have an updated second state based on the second detection region and the second state. generating an updated second tracking region based on the second state (Redden, [0022] , The confidence level for the point of interest is preferably increased if the point of interest is identified or extracted from a predetermined area or pixel-neighborhood (after accounting for movement of the system) in a subsequent image S134 (as shown in FIG. 3B), and preferably decreased otherwise. The confidence level for each point of interest is preferably updated with each new image,”); and
.
It would have been obvious to a person of ordinary skill before the effective filing date to modify Yang’s method in view of Redden to include a tracking area associated with the state of each target state in the sequence of images and update the tracking state based on second image state as taught by Redden with the benefit of detecting and tracking a target objects from the one or more targets from the plurality of targets based on a target state associated with each of the one or more target image (Redden , [0019]-[0022]).
Regarding claim 5, combination of Yang and Redden teaches the method according to claim 4.
Yang is silent on wherein updating the first tracker to have the updated second state comprises determining a union of the second detection region and the second tracking region to generate an updated second detection region and updating the first tracker based on the updated second detection region.
However, Qian teaches wherein updating the first tracker to have the updated second state comprises determining a union of the second detection region and the second tracking region to generate an updated second detection region and updating the first tracker based on the updated second detection region. (Redden, [0024] The portions of the sub-regions can overlap, wherein the overlapping region can be accounted for in both the first and the second sub-regions. However, the
foreground region can be segmented into sub-regions in any other suitable manner”).
It would have been obvious to a person of ordinary skill before the effective filing date to modify Yang’s method in view of Redden to include a tracking area associated with the state of each target state in the sequence of images and update the tracking state based on second image state as taught by Redden with the benefit of detecting and tracking a target objects from the one or more targets from the plurality of targets based on a target state associated with each of the one or more target image (Redden , [0019]-[0022]).
Regarding claim 6, combination of Yang and Redden teaches the method according to claim 4.
Yang is silent on wherein updating the first tracker to have the updated second state comprises matching the second tracking region to the second detection region based on a position of the second detection region relative to a position of at least one of: the second tracking region and another tracking region generated by the first tracker for another image of the sequence of images.
However, Redden teaches wherein updating the first tracker to have the updated second state comprises matching the second tracking region to the second detection region based on a position of the second detection region relative to a position of at least one of: the second tracking region and another tracking region generated by the first tracker for another image of the sequence of images (Redden,
).[0023] plant center repositioning input (e.g. the user moves the position of the plant center on the image), or any other suitable user input. Reclassifying points of interest as plant centers and non-plant centers S139 preferably includes
reclassifying the points of interest according to the received user input, wherein regions of the image indicated to be plant centers are reclassified as plant centers and regions of the image indicated to be non-plant centers are reclassified as non-plant centers. Reclassifying points of interest can
additionally include adding the user-edited image to the training set for the machine learning algorithms, wherein the reclassified plant centers can be used to better refine plant center identification, and the reclassified non-plant centers
can be used to better refine non-plant center classification. This step is preferably performed in near real-time, preferably before plant removal instruction generation
It would have been obvious to a person of ordinary skill before the effective filing date to modify Yang’s method in view of Redden to include a tracking area associated with the state of each target state in the sequence of images and update the tracking state based on second image state as taught by Redden with the benefit of detecting and tracking a target objects from the one or more targets from the plurality of targets based on a target state associated with each of the one or more target image (Redden , [0019]-[0022]).
Regarding claim 7, combination of of Yang and Redden teaches the method according to claim 6,
Yang is silent on wherein matching the second tracking region to the second detection region based on the first and second positions comprises matching the second tracking region to the second detection region based on a distance between a center of the second detection region and a center of the at least one of: the second tracking region and the another tracking region.
However, Redden teaches wherein matching the second tracking region to the second detection region based on the first and second positions comprises matching the second tracking region to the second detection region based on a distance between a center of the second detection region and a center of the at least one of: the second tracking region and the another tracking region. (Redden, [0019], “As shown in FIGS. 2 to 5, identifying individual plants includes identifying a foreground region
within the image Sll0, identifying points of interest within the foreground region S120, classifying the points of interest as plant centers or non-plant centers S130, and segmenting the foreground region into sub-regions S140, wherein each
sub-region encompasses a single point of interest classified as a plant center. The image from which the individual plants can be a frame of a video or a static image. Each image can be analyzed independently (diachronically) and/or in comparison to other images to account for changes in time ( e.g.synchronically). In addition to identifying individual plants, the plant bed area, the plant bed longitudinal axis, the distance between adjacent plant beds (e.g. rows), row locations, parallelism, and coherence ( e.g. how well plants are aligned within a given row) or any other suitable information about the plant beds and plant field can be estimated based on sensor data. This information can be used to
improve the computational performance of point of interest identification.).
It would have been obvious to a person of ordinary skill before the effective filing date to modify Yang’s method in view of Redden to include a tracking area associated with the state of each target state in the sequence of images and update the tracking state based on second image state as taught by Redden with the benefit of detecting and tracking a target objects from the one or more targets from the plurality of targets based on a target state associated with each of the one or more target image (Redden , [0019]-[0022]).
Regarding claim 8, combination of Yang and Redden teaches the method according to claim 7,
Yang is silent wherein the center of the second detection region comprises a prediction of the first plant's center of mass.
However, Redden teaches wherein the center of the second detection region comprises a prediction of the first plant's center of mass (Redden, [0019] [0019] Identifying individual plants Sl00 functions to distinguish individual plants within an image of contiguous, close-growing plants. More specifically, identifying individual plants functions to determine an identifying feature that is preferably only exhibited once in each plant. The identifying features are preferably plant centers,”).
It would have been obvious to a person of ordinary skill before the effective filing date to modify Yang’s method in view of Redden to include a tracking area associated with the state of each target state in the sequence of images and update the tracking state based on second image state as taught by Redden with the benefit of detecting and tracking a target objects from the one or more targets from the plurality of targets based on a target state associated with each of the one or more target image (Redden , [0019]-[0022]).
Regarding claim 9, combination of Yang and Redden teaches the method according to claim 7,
Yang is silent on wherein matching the second tracking region to the second detection region comprises generating a determination that the distance is less than at least a matching threshold distance and selecting the second detection region from a plurality of detection regions based on the determination.
However, Redden teaches (Redden,[0021] Classifying the points of interest can
include categorizing the points of interest into a plant center group or a non-plant group, can include assigning weights or confidence levels to each point of interest, indicative of the likelihood that said point of interest is a plant center, or can include any other suitable method or combination thereof of classifying the points of interest. While all points of interest within a given foreground region are preferably classified, a subset of points of interest can alternatively be classified,
particularly when the non-classified points of interest satisfy exclusion parameters ( e.g. the distance between the point of interest and the nearest edge is lower than a given percentage of the distance between the point of interest and the furthest edge, the point of interest area is below a size threshold,
etc.). In one variation, as shown in FIG. 2C, the points of interest can be classified using machine learning algorithms or artificial intelligence, wherein the machine learning or artificial intelligence algorithms are preferably supervised learning algorithms trained on a labeled set of examples ( e.g. images of plants with pre-identified plant centers) but can alternatively be unsupervised learning”).
It would have been obvious to a person of ordinary skill before the effective filing date to modify Yang’s method in view of Redden to include a tracking area associated with the state of each target state in the sequence of images and update the tracking state based on second image state as taught by Redden with the benefit of detecting and tracking a target objects from the one or more targets from the plurality of targets based on a target state associated with each of the one or more target image (Redden , [0019]-[0022]).
Regarding claim 10, combination of Yang and Redden teaches the method according to claim 1,
The method according to claim 1 further comprising:
detecting a second plant in a first second-plant (2P) image of the sequence of images (Yang, Figure 1-4, and 6, (Yang, Figure 6, [0039], “plant recognition engine 116 may be configured to obtain a digital image that captures at least a first plant of a plurality of plants”. Figure 3-4, 2303, 3361, 3362, and in figure 4, 4423 are time sequence images of a 3rd plant),
Yang is silent on by initializing a second tracker for the second plant based on the first 2P image, the second tracker having a first 2P state;
detecting the second plant in a second 2P image of the sequence of images by: updating the second tracker to have a second 2P state based on the first 2P state and the second 2P image; wherein the first 2P image comprises at least one of the first image, the second image, and a third image of the sequence of images, and the second 2P image comprises at least one of the second image, the third image, and a fourth image of the sequence of images, the second 2P image subsequent to the first 2P image in the sequence of images.
However, Redden teaches by initializing a second tracker for the second plant based on the first 2P image, the second tracker having a first 2P state;
detecting the second plant in a second 2P image of the sequence of images by: updating the second tracker to have a second 2P state based on the first 2P state and the second 2P image; wherein the first 2P image comprises at least one of the first image, the second image, and a third image of the sequence of images, and the second 2P image comprises at least one of the second image, the third image, and a fourth image of the sequence of images, the second 2P image subsequent to the first 2P image in the sequence of images. (Redden, [0007] FIGS. 3A, 3B, and 3C are schematic representations of a second variation of classifying the points of
interest as plant centers within the image, including identifying points of interest interest in a second image, and classifying recurring points of interest between the two images as plant centers, respectively [0024].in a first image, identifying points of at least a first, second, and third point on the foreground region edge proximal the respective point of interest, wherein the first edge point preferably opposes the second edge point across the respective point of interest, and the third edge point preferably opposes a point on the centerline across the respective point of interest. When the sub-regions are defined by ovals, the ovals preferably intersect the centerpoint and as many points on the foreground edge proximal the respective point of interest as possible”
It would have been obvious to a person of ordinary skill before the effective filing date to modify Yang’s method in view of Redden to include a tracking area associated with the state of each target state in the sequence of images and update the tracking state based on second image state as taught by Redden with the benefit of detecting and tracking a target objects from the one or more targets from the plurality of targets based on a target state associated with each of the one or more target image (Redden , [0019]-[0022]).
Regarding claim 11, combination of Yang and Redden teaches the method according to claim 10,
Yang further teaches wherein detecting the second plant in the first 2P image comprises: generating a 2P determination that less than a matching threshold area of at least one of: the first 2P tracking region and a first 2P detection region generated for the second plant overlaps with each of at least one of: one or more detection regions and one or more tracking regions for one or more other plants, the one or more other plants comprising at least the first plant; and validating the at least one of: the first 2P tracking region and a first 2P detection region based on the 2P determination. (Yang, Figure 3, a bounding shape may be detected that captures, for instance, some predetermined percentage of the plant, or a portion of the plant that is identified with at least a threshold amount of confidence. For example, the tips of
leaves of a first plant may overlap with a neighboring plant. Accordingly, those overlapping portions of the first plant may not be identified as being part of the first plant with as much confidence as the middle of the first plant, and
therefore may not necessarily be captured by a bounding shape. Similar to first bounding shape 3361, a second bounding shape 3362 is detected around third plant 2303 at twenty-five days. As is depicted in FIG. 3, second bounding
shape 3362 is larger than first bounding shape 3361. This is unsurprising given that third plant 2303 grew during this time interval.).
Regarding claim 14, Yang and Redden teaches the method according to claim 1
Yang further teaches wherein: detecting one or more plants in a first image comprises detecting up to n plants in the sequences of images, the up to n plants comprising the first plant, for a predetermined n; (Yang, Figure 6, [0039], “plant recognition engine 116 may be configured to obtain a digital image that captures at least a first plant of a plurality of plants”. Figure 3, time sequence images” NOTE: the plant recognition engine 116 obtain images from plurality of plants. Therefore, there could be any number of detection region from 1-n. See [0005]” In some implementations, a unique identifier may be generated for each unique plant. A plurality of plants, e.g.,of a field or farm, may be indexed in a database using their respective unique identifiers”.)
generating the first detection region comprises: generating at least n + 1 detection regions based on the sequence of images; and selecting up to n detection regions of the at least n + 1 detection regions; (Yang, Figure 3-4, [0063],”as depicted in FIG. 3, the latent space embeddings 446 1_2 generated from the anchor image 4421 and positive image 4422-both depicting the same plant 4301-are fairly similar, and therefore should be close together in latent space”. NOTE 446 is interpreted as detection region for the first plant based on first image. True for n number of detection regions.);
Yang is silent on initializing the first tracker comprises initializing up to n trackers, each of the up to n trackers comprising a corresponding tracking region based on a corresponding one of the up to n detection regions; and the up to n plants comprise the first plant, the up to n detection regions comprise the first detection region, and the up to n trackers comprise the first tracker.
However, Redden teaches initializing the first tracker comprises initializing up to n trackers, each of the up to n trackers comprising a corresponding tracking region based on a corresponding one of the up to n detection regions; and the up to n plants comprise the first plant. (Redden, [0025] “Identifying individual plants can additionally include creating a virtual map of plants S160, which functions to
map the sub-regions or regions indicative of plants that are extracted from the image to a virtual map of the plant field. As shown in FIG. 7A, the virtual map functions to identify the relative position of each plant within the field, and can
also function to identify the relative position of the crop thinning system relative to each identified plant. The virtual map preferably correlates directly with the actual positions of the plants ( e.g. is a 1: 1 representation of the analyzed
portion of the plant field), but can alternatively be any other suitable representation of the actual plants on the field. As shown in FIG. 7B, the virtual map is preferably dynamically updated S162 and/or expanded with each successive image. Each sub-region or region encompassing a single point of
interest classified as a plant center is preferably treated as a plant, and is preferably mapped to a position within the virtual map.)
It would have been obvious to a person of ordinary skill before the effective filing date to modify Yang’s method in view of Redden to include a tracking area associated with the state of each target state in the sequence of images and update the tracking state based on second image state as taught by Redden with the benefit of detecting and tracking a target objects from the one or more targets from the plurality of targets based on a target state associated with each of the one or more target image (Redden , [0019]-[0022]).
Regarding claim 15, combination of Yang and Redden teaches the method according to claim 14,
Yang further teaches determining, for each of the n + 1 detection regions (Yang, Figure 6, [0039], “plant recognition engine 116 may be configured to obtain a digital image that captures at least a first plant of a plurality of plants”. Figure 3, time sequence images):,
wherein selecting up to n selected detection regions comprises an associated probability that the detection region contains a plant based on trained parameters of a machine learning model determining, for each of the n selected detection regions, that the associated probability for the selected detection region is at least as great as the associated probability for each of the n + 1 detection regions not in the up to n selected detection regions; and selecting up to n of the n + 1 detection regions based on the determining, for each of the n selected detection regions, that the associated probability for the selected detection region is at least as great as the associated probability for each of the n + 1 detection regions not in the up to n selected detection regions (Yang,[0007] Various types of machine learning models may be trained and utilized to recognize individual plants using image data. In some implementations, a convolutional neural network ("CNN") may trained using techniques such as triplet loss to recognize individual plants. In other implementations, a sequence-to-sequence model such as an encoder-decoder may be trained to recognize individual plants, e.g., by using a sequence of images known to depict a particular plant to predict the plant's appearance in a next image and matching that predicted appearance to a ground truth image obtained in the field”) ;
.
Regarding claim 16, combination of Yang and Redden teaches the method according to claim 14,
Yang further teaches wherein selecting up to n selected detection regions comprises:
for each of a first and second candidate detection region, (Yang, Figure 6, [0039], “plant recognition engine 116 may be configured to obtain a digital image that captures at least a first plant of a plurality of plants”. Figure 3, time sequence images):,
Yang is silent on determining a spectral characteristic value based on a corresponding portion of the first image; and selecting the first candidate detection region and rejecting the second candidate detection region based on a comparison of the spectral characteristic value for the first candidate detection region and the spectral characteristic value for the second candidate detection region.
However, Redden teaches determining a spectral characteristic value based on a corresponding portion of the first image; and selecting the first candidate detection region and rejecting the second candidate detection region based on a comparison of the spectral characteristic value for the first candidate detection region and the spectral characteristic value for the second candidate detection region. (Redden, [0026] Selecting sub-regions from the image preferably includes determining an optimal pattern of retained plants that best meets a set of cultivation parameters for the section of the plant field represented by the image, based on the parameters of each sub-region, and selecting the sub-regions within the image that correspond to the retained plants to obtain the optimal retained plant pattern. The cultivation
parameters can include the intra-row distance between adjacent plants (e.g. distance between plants in the same row), inter-row distance between adjacent plants ( e.g. distance between plants in different rows), yield (e.g. number of
plants per unit area or plant density), uniformity in plant size (e.g. uniformity between the retained sub-regions and/or foreground region area), uniformity in plant shape (e.g. uniformity in retained sub-region and/or foreground region
perimeter), uniformity in plant appearance, plant size and/or shape similarity to a given size and/or shape, the confidence or probability that the sub-region or region is a plant, the practicality of keeping the respective plant, uniformity or
conformance of measured plant health indicators to a plant health indicator threshold, or any other suitable parameter that affects the plant yield that can be determined from the information extracted from the imag”)
It would have been obvious to a person of ordinary skill before the effective filing date to modify Yang’s method in view of Redden to include a tracking area associated with the state of each target state in the sequence of images and update the tracking state based on second image state as taught by Redden with the benefit of detecting and tracking a target objects from the one or more targets from the plurality of targets based on a target state associated with each of the one or more target image (Redden , [0019]-[0022]).
Regarding claim 17, combination of Yang and Redden teaches the method according to claim 16,
Yang is silent on wherein: the spectral characteristic value for at least the first candidate detection region is based on a histogram distance between the corresponding portion of the first image for the first candidate detection region and one or more corresponding portions of the first image for one or more other ones of the n + 1 detection regions; and selecting the first candidate detection region comprises determining that the spectral characteristic value for the first candidate detection region is less than the spectral characteristic value for the second candidate detection region
However, Redden teaches wherein: the spectral characteristic value for at least the first candidate detection region is based on a histogram distance between the corresponding portion of the first image for the first candidate detection region and one or more corresponding portions of the first image for one or more other ones of the n + 1 detection regions; and selecting the first candidate detection region comprises determining that the spectral characteristic value for the first candidate detection region is less than the spectral characteristic value for the second candidate detection region (Redden, [0021], “Classifying the points of interest can include categorizing the points of interest into a plant center group or a non-plant group, can include assigning weights or confidence levels to each point of interest, indicative of the
likelihood that said point of interest is a plant center, or can include any other suitable method or combination thereof of classifying the points of interest. While all points of interest within a given foreground region are preferably classified, a
subset of points of interest can alternatively be classified, particularly when the non-classified points of interest satisfy exclusion parameters ( e.g. the distance between the point of interest and the nearest edge is lower than a given percentage of the distance between the point of interest and the furthest edge, the point of interest area is below a size threshold, etc.).”).
It would have been obvious to a person of ordinary skill before the effective filing date to modify Yang’s method in view of Redden to include a tracking area associated with the state of each target state in the sequence of images and update the tracking state based on second image state as taught by Redden with the benefit of detecting and tracking a target objects from the one or more targets from the plurality of targets based on a target state associated with each of the one or more target image (Redden , [0019]-[0022]).
Regarding claim 18, combination of Yang and Redden teaches the method according to claim 1,
Yang further teaches wherein generating a first detection region comprises:
extracting a plant mask based on trained parameters of a machine learning model, the plant mask mapping regions of the first image to probabilities that the regions include a plant; identifying one or more objects in the first image based on the plant mask; generating the first detection region for the first plant based on at least one of the one or more objects in the plant mask. (Yang, 0007] Various types of machine learning models may be trained and utilized to recognize individual plants using
image data. In some implementations, a convolutional neural network ("CNN") may trained using techniques such as triplet loss to recognize individual plants. In other implementations, a sequence-to-sequence model such as an
encoder-decoder may be trained to recognize individual plants, e.g., by using a sequence of images known to depict a particular plant to predict the plant's appearance in a next image and matching that predicted appearance to a ground
truth image obtained in the field”).
Regarding claim 19, combination of Yang and Redden teaches the method according to claim 18,
Yang further teaches wherein the first detection region comprises a bounding box (Yang, Figure 3, [0041] This additional data may take other forms in addition
to or instead of position coordinates. For example, in various implementations, a bounding shape may be calculated for the first plaint. Various techniques such as edge detection, machine learning (e.g., a convolutional neural network, or "CNN"), segmentation, etc., may be employed to detect a bounding shape that encloses at least a portion of the first plant”).
Regarding claim 20, combination of Yang and Redden teaches the method according to claim 18,
Yang further teaches wherein extracting the plant mask comprises extracting a background mask based on the trained parameters of the machine learning model, the background mask mapping regions of the first image to probabilities that the regions include non-plant background; and generating the plant mask based on an inversion of the background mask (Yang, Figure 2-4,[0057] In various implementations, various aspects of bounding shapes 3361_2 may be used as additional plant attributes to recognize third plant 2303 as distinct from other plants 2301_2 and 2304 _5 . For example, a width and/or height of bounding shapes 3361_2 may be used as proxies for dimensions of third plant 2303 . These additional attributes may be applied, e.g., by plant recognition engine 116, as additional inputs to a machine learning model, along with the digital images in which the bounding shapes 3361_2 were detected. The output of the machine learning model may distinguish the plant from other plants based at least in part on these additional attributes”). .
Regarding claim 21, combination of Yang and Redden teaches the method according to claim 1,
Yang further teaches wherein the method comprises estimating at least one of: a diameter, a biomass, and a height of the first plant based on at least one of: the first tracking region and the second tracking region for the first plant (Yang, [0011] Various aspects of bounding shapes that enclose individual plants may also be used, in addition to or instead of other plant attributes described herein, as additional
attributes of those plants. For example, an image of a plant may be processed to determine a minimum bounding shape (e.g., square, rectangle, circle, polygon, etc.) that encloses at least a portion of the plant (e.g., stems and fruit), if not the
entire plant. Various aspects of that bounding shape, such as a diameter, radius, width, height, number of edges used (in the case of polygon bounding shapes), may be used as additional attributes of the individual plant”) .
Reasons for Allowability / Allowable Subject Matter
Claims 12-13 are allowed.
The following is an examiner's statement of reasons for allowance:
Regarding Claim 12, the prior art of record as considered and understood by the examiner fails to teach or fairly suggest:
“generating a third detection region generating a 3P determination that more than the matching threshold area of the third detection region overlaps with at least one of: the first detection region, the second detection region, the first 2P detection region, and the second 2P detection region; and invalidating the third detection region based on the 3P determination.
updating the second tracker to have a second 2P state based on the first 2P state and the second 2P image.”
in combination with all other claimed limitations of claim 12.
Furthermore, there is not any obvious motivation for an ordinary skilled in the art
to combine some and/or all of the features of the prior art of record to achieve the
features of the independent claim. In other words, it will further require substantial structural modification of the components that will also require substantial modification of the measurements and data processing to achieve the features of the allowable subject matter set forth in the independent claims.
Regarding Claim 13, the prior art of record as considered and understood by the examiner fails to teach or fairly suggest:
“wherein the matching threshold area comprises 50% of an area of at least one of: the first 2P tracking region, the first 2P detection region, any of the one or more detection regions for the one or more other plants, and any of the one or more tracking regions for one or more other plants .”
in combination with all other claimed limitations of claim 12.
Furthermore, there is not any obvious motivation for an ordinary skilled in the art
to combine some and/or all of the features of the prior art of record to achieve the
features of the independent claim. In other words, it will further require substantial structural modification of the components that will also require substantial modification of the measurements and data processing to achieve the features of the allowable subject matter set forth in the independent claims.
Conclusion
Citation of Pertinent Prior Art
The prior art made of record and not relied upon is considered pertinent to
applicant's disclosure.
Binney et al. (US 2022/0007589 A1) recites “One variation of a method for monitoring growth of plants within a facility includes: aggregating global ambient data
recorded by a suite of fixed sensors, arranged proximal a grow area within the facility, at a first frequency during a grow period; extracting intermediate outcomes of a set of
plants, occupying a module in the grow area, from module level images recorded by a mover at a second frequency less than the first frequency while interfacing with the module during the period of time; dispatching the mover to autonomously
deliver the module to a transfer station; extracting intermediate outcomes of the set of plants from plant-level images recorded by the transfer station while sequentially
transferring plants out of the module at the conclusion of the grow period; and deriving relationships between ambient conditions, intermediate outcomes, and final outcomes from a corpus of plant records associated with plants grown in the facility.” (abstract).
Shinji Watanabe (US 2018/0268557 A1) discloses “According to some aspects, an information processing device is provided. The information processing device includes circuitry configured to dispose a plurality of tracking points within a first region of a first image and set a second region of a second image based on estimated positions of the plurality of tracking points in the second image. The estimated positions are determined by comparing the first image and the second image. The circuitry is further configured to re-dispose the plurality of tracking points within the second region of the second image.
” (Abstract):
Lipchin et al. (US 2020/0097769 A1) recites “Methods, systems, and techniques for object detection and tracking are provided. A system may include a module configured to generate a plurality of region proposals, each region proposal comprising a part of a video frame, a CNN pre-trained for object detection, the plurality of region proposals being input to the CNN; a tracker for tracking one or more targets based on outputs from the CNN across the series of video frames and generating tracking information on the one or more targets; and a module further configured to refine the plurality of region proposals to be input to the CNN, based on the tracking information” (abstract).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DILARA SULTANA whose telephone number is (571)272-3861. The examiner can normally be reached Mon-Fri, 9:00AM-6 PM.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, EMAN ALKAFAWI can be reached on (571) 272-4448. 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.
/DILARA SULTANA/Examiner, Art Unit 2858
05/30/2026
/EMAN A ALKAFAWI/Supervisory Patent Examiner, Art Unit 2858 6/5/2026
Prosecution Insights