DETAILED ACTION
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 Remarks
Claim Objections
The objection to the claim is withdrawn in light of Applicant’s amendments
Claim Rejections - 35 USC § 112(b)
Applicant's arguments filed 12/29/2025 have been fully considered but they are only partly persuasive. While the claim is cancelled and thus the rejection of Claim 8 withdrawn, the same general limitations at issue are found in new Claims 24, 26, and 27. No rejections beyond cancellation of the original Claim 8 such as addressing the issue generally recreated in the new claims is provided. See the updated rejection below.
Claim Rejections - 35 USC § 102
Applicant's arguments filed 12/29/2025 have been fully considered but they are not persuasive.
First, the language of the claims is not as narrow as Applicant has argued. Applicant argues “two separate sets of joint positions” and “different robotic devices”, however at minimum the independent claims are not this specific. Applicant merely uses the numerical adjectives of “first”, “second”, and “third” which merely provide for an easy way to refer to potentially differing items. There is no inherent requirement for the items to actually differ. Additionally, these adjectives do not inherently provide for what might even make them different. Furthermore, Applicant’s disclosure appears silent with respect to simultaneous or even concurrent or following mapping, let alone control outputs (see related discussion below), while instead making clear that particular items may or may not be the same despite use of the “first” and “second” adjectives. See for example [0047] “The first set of joint positions and the second set of joint positions may comprise a same or different number of joints”, [0061] “the term body plan used herein refers to a number of legs for a body”, and [0053] “the virtual body and the robotic device may have a same body plan” of Applicant’s originally filed specification. Therefore, the broadest reasonable interpretation of the claim terms, particularly in light of the Applicant’s explicit disclosure, is much broader than argued.
Additionally, Applicant appears to correlate required limitations of the “set of joint positions” not presently required. The claims presently only provide inputting to the machine learning model but do not provide any clear limitations of outputs thereof, only the nature of the way the model has been trained.
Third, the cited references disclose these limitations, even in light of more narrow interpretations than appears appropriate (see above) or supported (see below). See the updated rejection below.
Furthermore, and relatedly, the appeared intended meaning of the limitations based on Applicant’s argument is that the machine learning model is not just trained to be capable of mapping a given set of joint positions to one other body but at least two bodies, but furthermore that the system processes and outputs control information for more than one body in a given execution/process/method. This does not appear support by Applicant’s originally filed disclosure. See the 112(a) Section below for the related rejections of the claims.
If Applicant instead only meant that the machine learning model has a particular capacity, the language of the claim should be altered such that only one “outputting” action is/was recited, possibly using an alternative “or” statement or similar to refer to the joint positions of interest.
Claim Rejections - 35 USC § 103
Applicant’s arguments appear to rely solely on those made with respect to the independent claim rejections under 35 USC § 102 which are already addressed above.
Claim Interpretation
Recitations of “to [verb]” have been interpreted as functional limitations meaning “configured to” such that a particular structure is required.
Applicant makes frequent use of the phrase “for” which typically indicates a purpose, intended result, or similar and not a particular functional or structural limitation. While Examiner has addressed many “for” limitations under prior art as if positively recited functional or structural limitations in the interest of compact prosecution as such features were already disclosed in the prior art, Examiner notes that the broadest reasonable interpretation may be inclusive of mere purpose, intended result, or similar.
Claim Rejections - 35 USC § 112(a)
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claims 14 – 33 are rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention.
Regarding Claims 14, 15, and 33, the claims recite the limitation:
“outputting control information to the second robotic device in dependence on the third set of joint positions for controlling one or more actuators of the second robotic device”
following the preceding limitation of:
“outputting control information to the first robotic device in dependence on the second set of joint positions for controlling one or more actuators of the first robotic device”.
Applicant has not pointed out where the amended and/or new claim is supported, nor does there appear to be a written description of the above claim limitation in the application as filed. MPEP 2163 relates which states “Applicant should ... specifically point out the support for any amendments made to the disclosure”. Examiner has made a best effort to identify support for the claims and in particular the identified limitation above without success.
The closest support in Applicant’s originally filed specification was found in:
[0056] which indicates that a particular “target number of joints” that the model may be trained for may be “associated with one or more robotic devices that are intended to be used”. However, this merely indicates what it states; that a particular “set of joint positions” may be associated with more than a single robotic device. It does not indicate simultaneous or even concurrent or consecutive mapping within a same process, and is unrelated to control outputs which are indicated as being based on the machine learning outputs, not an output of the machine learning itself (see e.g. [0049]).
[0067], which appears to be the closest to support for the limitations but which appears to only indicate that a machine learning model might be trained using more than two body types for a given “same type of configuration”.
Neither of these paragraphs indicate simultaneous or even concurrent or consecutive mapping within a same process, and are unrelated to control outputs which are indicated as being based on the machine learning outputs, not an output of the machine learning itself (see e.g. [0049]). They at most appear to indicate that the output may be mapped to one of a plurality of body types and that a given mapping may correlate to more than one robot (more than one robot being associated with the target number of joints of the output). See e.g. [0049] which repeatedly makes clear that the output is for a singular robotic device and the drawings which do not disclose a plurality of outputs or robots from a given input.
Therefore, the claim contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for pre-AIA the inventor(s), at the time the application was filed, had possession of the claimed invention.
Regarding Claims 16, 18 – 19, and 21, the claims recite features which build off the of the features indicated above as lacking supported and are consequently rejected under the same logic. The closest support is already recited and clearly lacks the detail provided in the claim, let alone the features from which they depend.
Regarding Claims 17, 20, and 22 – 32, the claims depend from claim(s) rejected above and inherit the deficiencies of said claim(s) as described above. Therefore, Claims 17, 20, and 22 – 32 are rejected under the same logic presented above.
Claim Rejections - 35 USC § 112(b)
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim 27 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Regarding Claim 27, the claim recites the limitation “wherein the first set of training joint positions and the second set of training joint positions comprise joint positions for one or more body configuration types selected from the list consisting of”. This limitation is preceded by and appears intended to further limit the limitation of Claim 24 which recites “training data comprising sets of joint positions for a plurality of same body configuration types for each of a first body having a first number of legs and a second body having a second number of legs”.
It is unclear if the “set of training joint positions” of Claim 27 are the same or in addition to those “sets of joint positions” of Claim 24. Note that Claim 26 states that these are part of “the training data” which comprises “sets of joint positions”, however Claim 26 does not indicate that these are of the “sets of joint positions of Claim 24. Furthermore, Claim 24 joint positions are “of same body configuration types for each of a first body and a second body” indicating there are at least first and second sets.
Furthermore, Claim 27 is a Markush claim reciting a closed list of alternatives (see “list consisting of”). Therefore, it is highly unclear exactly how limited the “sets of joint positions” are, as the higher-level language for the claims uses “consisting” and therefore even if a body configuration type was outside of the closed Markush list, it could simply be interpreted as a different body configuration type of a different “set of joint positions” effectively rendering the Markush claim language moot, being no more limiting than a “comprising” limitation.
In the interest of compact prosecution, Claim 27 has been interpreted as instead reading “wherein the sets of joint positions are for one or more body configuration types selected from the list consisting of”.
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claims 14 – 15, 22, 24 – 25, 29 – 33 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Kim et al. (Kim, Sunwoo, et al. "Human Motion Control of Quadrupedal Robots using Deep Reinforcement Learning." arXiv preprint arXiv:2204.13336 (2022)).
Regarding Claim 14, Kim teaches:
A method comprising:
generating images for display comprising a virtual body comprising one or more virtual limbs (See at least Section I, “We demonstrate that our system allows a human user to execute various motor tasks with simulated and real quadrupedal robots using a consumer-grade motion capture system, Microsoft Kinect [45]”, Section III, “Our system receives human motions from any motion capture system, which is Microsoft Azure Kinect [45] in our case. Then the motion retargeting module (Section IV) converts the captured human into the corresponding robot reference motion that is physically valid and conveys proper semantics”, as well as Figure 2, Figure 5, and Figure 8 which illustrate a purple skeleton that is virtual corresponding to the RGB-D Video footage, or in other words demonstrates a virtual body having a plurality of limbs in an image format generated using well-known and understood motion capture hardware and techniques);
generating virtual body configuration information indicative of one or more first sets of joint positions for the virtual body in one or more of the images (See at least again above citations. The claim does not specify the definition of “virtual body configuration information”; the phrase “indicative of” is highly broad and non-specific, and the images clearly require joint information. Additionally, see at least Section III, Subsection A “We aim to develop a motion mapper f takes a human pose q as inputs and maps it into the corresponding robot pose p” and Figure 3 and associated caption); and
inputting the virtual body configuration information to a machine learning model trained to map at least one first set of joint positions for the virtual body to a second set of joint positions for a first robotic device comprising a plurality of first joints (See again at least above citations, in particular Figure 3 with “Motion Retargeting Network”) and to a third set of joint positions for a second robotic device comprising a plurality of second joints (See at least Section I “We demonstrate that our system allows a human user to execute various motor tasks with simulated and real quadrupedal robots”. Alternatively, see Figure 2 and Figure 5, and “Motion Imitation” of Section V. wherein machine learning is used again. Applicant does not limit, define, or otherwise describe the nature of their machine learning model),
outputting control information to the first robotic device in dependence on the second set of joint positions for controlling one or more actuators of the first robotic device (See at least Figure 2, final box of “Real-world Robot Motions” and various discussion of a controller and control, for example Section V “Because a human cannot reproduce the exact same motion, our controller must be able to imitate similar motions with spatial and temporal noises”, Figure 2, “Retargeted Robot Motions” and Figure 2 “Motion Imitation”); and
outputting control information to the second robotic device in dependence on the third set of joint positions for controlling one or more actuators of the second robotic device (See at least Figure 2, final box of “Real-world Robot Motions” and various discussion of a controller and control, for example Section V “Because a human cannot reproduce the exact same motion, our controller must be able to imitate similar motions with spatial and temporal noises”, Figure 2, “Retargeted Robot Motions” and Figure 2 “Motion Imitation”).
Regarding Claim 22, Kim teaches:
The method of claim 14, wherein the virtual body is a two-legged body and the first robotic device is a four-legged robotic device (See at least Figure 1 and Figure 2).
Regarding Claim 24, Kim teaches:
The method of claim 14, wherein the machine learning model is trained using training data comprising sets of joint positions for a plurality of same body configuration types for each of a first body having a first number of legs and a second body having a second number of legs (See at least Section IV, Subsection A, “Data preparation. We prepare the dataset D by collecting matching pairs of human and robot motions. First, we generate robot motions for sampled tasks. For example, tilting or manipulation tasks are synthesized by generating random goals and solving robot poses using inverse kinematics … Once we have the robot motions, we collect the matching human motion sequences. While showing robot motions, we ask a human user to act the “corresponding motions” based on the user’s own intuition and record the motions using a motion capture system … We use multi-layer perceptron (MLP) for training a mapper from the given dataset D (Figure 3)”
Furthermore, in the interest of compact prosecution, Examiner notes that the presently provided claim language does not specify that the first body and second body or first number of legs and second number of legs must be different).
Regarding Claim 25, Kim teaches:
The method of claim 24, wherein the first body is a two-legged body and the second body is a four-legged body (See at least Figure 1 and Figure 2).
Regarding Claim 29, Kim teaches:
The method of claim 14, wherein the images for display are images for a video game (Examiner notes that the intent of the images (see use of “for”) does not appear to change the nature of the images. They are structurally and functionally the same regardless of what they are for or how they are used as presently claimed. In the interest of compact prosecution, see also [0002] and [0014] of Cashman et al. (US 20230326135 A1) cited elsewhere which discloses use of such images for video games which may be relied upon should Applicant amend the claim such that more than just an intent or purpose is described).
Regarding Claim 30, Kim teaches:
The method of claim 14, further comprising controlling the virtual body using user inputs (See at least Figure 2, Figure 5, and discussion of “motion capture”).
Regarding Claim 31, Kim teaches:
The method of claim 14, further comprising inputting speed information (Examiner notes that per Applicant’s specification and the immediate depending claim, Claim 32, speed information may simply mean time series position information or similar) associated with movements by the virtual body to the machine learning model to control a speed of one or more actions by the first robotic device (See at least Figure 2 and Figure 5, caption of Figure 5 stating “at the corresponding time frames”, Section IV, Subsection A, “We further manually process the motions to clean up noisy segments and fix asynchronous actions based on contact flags. Once we obtain the motions, we compute the pose derivatives for both the user and the robot using finite differences with ∆t = 0.1”).
Regarding Claim 32, Kim teaches:
The method of claim 31, wherein the speed information comprises time information that characterizes a time associated with one or more joint positions of the virtual body (See again Claim 31).
Regarding Claims 15 and 33, the claims are directed to effectively the same subject matter as Claim 14 with respect to the application of prior art. The claims are therefore rejected under the same logic as Claim 14 above. Applicant’s Remarks filed 12/29/2025 which argue the independent claims collectively support this position.
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.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 16 – 18 are rejected under 35 U.S.C. 103 as being unpatentable over Kim et al. in view of Tao et al. (US 20240331248 A1) and Cashman et al. (US 20230326135 A1).
Regarding Claim 14, another approach to Claim 14 is that of simultaneously, iteratively, concurrently, or similarly generating joint positions for various desired end-outputs. This combination of Kim and Tao would change the mapping of Kim such that it may generate any number of outputs as required by iteration. See at least [0035] “Then, at block 316, the process may determine if additional iterations are to be performed. If so, the process may return to block 302 for the next iteration. Otherwise, the process may stop and/or the machine learning retargeting model may be utilized for normal operations of retargeting existing animations for different skeletons and/or different geometries” and Figure 3.
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to iterate the process as taught by Tao in Kim with a reasonable expectation of success. It is well understood and routine, if not simply common knowledge, to iterate a process required to be performed on/for different targets, subjects, etc.
Regarding Claim 16, Kim teaches:
The method of claim 14,
Kim does not teach, but Cashman teaches in the context of the combination of Kim and Tao:
wherein the first robotic device comprises a first number of first joints and the second robotic device comprises a second number of second joints, the first number of first joints being different from the second number of second joints (See at least [0002] “For example, different users may choose to be represented in the virtual environment with different virtual representations that have different proportions and/or a different number of joints from the training representation” and [0036] “For instance, the target articulated representation may have any suitable appearance and proportions, and may have any suitable number of limbs, joints, and/or other moveable body parts”).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention for the iterated robots to differ as disclosed by Cashman in the system and method of the combination of Kim and Tao. There would be minimal utility to iterate the same input across the same output. As disclosed by at least Tao, the purpose of such a machine learned process is to “allow an existing animation to be repurposed for a different skeleton and/or a different environment geometry from that associated with the existing animation”. The combination above allows for flexibility and repeatability in application, a feature also appearing at least implied in Kim (note the repeated use of the word “robots”).
Regarding Claim 17, the combination of Kim, Cashman and Tao teaches:
The method of claim 16,
Kim further teaches:
wherein the virtual body comprises a third number of joints that is different from the first number of first joints (See at least [0039] “In the above example, there is a one-to-one correspondence between joints in the target articulated representation and joints in the model articulated representation. It will be understood, however, that this need not always be the case. Rather, in some examples, two or more different joints in the model articulated representation may map to a single joint in the target articulated representation. For example, the target articulated representation may include fewer spinal joints than the model articulated representation. Thus, a joint mapping constraint may specify that two or more different spinal joints in the model representation influence one spinal joint in the target representation”).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention for the number of joints between the source body and the retargeted body to be different as taught by Cashman in the system and method of Kim or Kim in combination with Tao with a reasonable expectation of success. The body of a bipedal humanoid and animal-like quadruped of Kim are likely different including that of the number of joints involved, particularly as one is a robot rather than live creature. As necessary, the system of Cashman would account for such differences, and at minimum discloses that such a distinction is common and expected.
Regarding Claim 18, the combination of Kim, Cashman and Tao teaches:
The method of claim 17,
Regarding the following limitation, the same logic and teachings appear to apply as with respect to Claim 17, wherein the disclosure as a whole is directed towards retargeting a particular motion between one source and varying target skeletons/topologies/morphologies/rigs/etc.
wherein the third number of joints is different from the second number of second joints.
Claims 19 – 21 are rejected under 35 U.S.C. 103 as being unpatentable over Kim et al. in view of Payne et al. (US 20200394806 A1) and Cashman et al.
See again note regarding combination of Kim and Tao above.
Regarding Claim 19, Kim teaches:
The method of claim 14,
Kim does not teach, but Payne teaches in the context of the combination of Kim and Tao:
wherein the first robotic device comprises a first number of legs and the second robotic device comprises a second number of legs, the first number of legs being different from the second number of legs (See at least [0040] “As another example, each animated character may have a unique body shape (e.g., they may be associated with different types of animals), such as different numbers of limbs, different limb lengths or designs, and so forth. These differences may result in each animated character having a distinct version of a specific pose” and Figure 2 showing various target characters).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention for the iterated robots to differ as disclosed by Payne in the system and method of the combination of Kim and Tao. There would be minimal utility to iterate the same input across the same output. As disclosed by at least Tao, the purpose of such a machine learned process is to “allow an existing animation to be repurposed for a different skeleton and/or a different environment geometry from that associated with the existing animation”. The combination above allows for flexibility and repeatability in application, a feature also appearing at least implied in Kim (note the repeated use of the word “robots”). See also [0002] of Payne which discloses the benefits an iterative process reusing existing resources provides.
Regarding Claim 20, the combination of Kim, Payne, and Tao teaches:
The method of claim 19,
Kim further teaches:
wherein the virtual body comprises a third number of legs that is different from the first number of legs (See at least Figure 1).
Regarding Claim 21, the combination of Kim, Payne, and Tao teaches:
The method of claim 20,
Regarding the following limitation, the same logic and teachings appear to apply as with respect to Claim 19, wherein the disclosure as a whole is directed towards retargeting a particular motion between one source and varying target skeletons/topologies/morphologies/rigs/etc. even going so far as to disclose a “universal biomechanical expression system” in Payne intended to streamline the system such that a given input may be applied across a variety of highly differing outputs.
wherein the third number of legs is different from the second number of legs.
Claim 23 is rejected under 35 U.S.C. 103 as being unpatentable over Kim et al. in view of Hu et al. (Hu, Lei, et al. "Pose-aware Attention Network for Flexible Motion Retargeting by Body Part." arXiv preprint arXiv:2306.08006 (2023)).
Regarding Claim 23, Kim teaches:
The method of claim 14,
Kim does not teach, but in combination with Hu teaches:
wherein the virtual body is a four-legged body and the first robotic device is a two-legged robotic device (See at least Figure 11, “Qualitative results of retargeting between bipedal and quadrupedal”).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to utilize the flexible motion retargeting process of Hu, including that of a bipedal source to quadrupedal target, in the system and method of Kim with a reasonable expectation of success. Hu is designed to be flexible such that problems “such as the source-target skeletons needing to have the same number of joints or share the same topology” (Abstract of Hu) are overcome. Hu states that “Extensive experiments show that our approach can generate better motion retargeting results both qualitatively and quantitatively than state-of-the-art methods. Moreover, we also show that our framework can generate reasonable results even for a more challenging retargeting scenario, like retargeting between bipedal and quadrupedal skeletons because of the body part retargeting strategy and PAN”.
Claims 26 – 28 are rejected under 35 U.S.C. 103 as being unpatentable over Kim et al. in view of Cashman et al. (US 20230326135 A1).
Regarding Claim 26, Kim teaches:
The method of claim 24,
Kim does not explicitly teach, but Cashman teaches:
wherein the training data comprises, for a given body configuration type:
a first set of joint positions for the first body and associated with a given label corresponding to the given body configuration type; and
a second set of joint positions for the second body and associated with the given label (See at least [0014] “In some cases, the pose of the virtual articulated representation may be output by a previously-trained model, which may include a neural network as a non-limiting example. As discussed above, the model may in some cases be trained with training data having ground truth labels for the same virtual articulated representation that is displayed in the virtual environment. For instance, the training data may include positioning data for a human training subject's body parts (e.g., rotation parameters corresponding to the subject's joints), labeled with ground truth labels relating the positioning data to a pose of the corresponding joints of the virtual articulated representation that matches the subject's real-world pose. Based on such training data, the model may output poses for the virtual articulated representation for any suitable set of input positioning data”, [0063] “Furthermore, pose estimation may in some cases be performed by a previously-trained machine learning (ML) model 508 implemented by the pose optimization machine. As discussed above, the previously-trained pose optimization machine is trained with positioning data having ground truth labels for the model articulated representation. For example, the pose optimization machine may implement a neural network or other suitable model that is previously-trained by providing training positioning data specifying rotation parameters for joints of one or more training subjects, and the ground truth labels may specify a pose of the model articulated representation that would be consistent with the real-world pose of the training subject”, and “[0066] Furthermore, in some cases, the training positioning data for the pose optimization machine may not include ground truth labels for the target articulated representation” (“may not”, particularly following “furthermore” means that the converse is also true”).
Examiner additionally notes that the claims use terms and phrases which have particularly broad meanings. For example, the claim uses the phrase “corresponding to” without actually describing or claiming the nature of said correspondence or uses terms such as “label” or “body configuration type” without defining the terms, and wherein Applicant’s specification provides neither a special definition or particular limit to the interpretation of the terms).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention for the training data to be labeled as disclosed by Cashman in the system of Kim with a reasonable expectation of success. Such labeling and use of ground truth data is typical in training machine learning models and serves to ensure desired correlations are being taught to the model. In the case of Kim and Cashman, the source and retargeted bodies, and mapping from/to thereof are the desired correlations.
Regarding Claim 27, the combination of Kim and Cashman teaches:
The method of claim 26,
Kim further teaches:
wherein the first set of training joint positions and the second set of training joint positions comprrise joint positions for one or more body configuration types selected from the list consisting of (i) a neutral standing configuration (See at least Section V “First, we design a hierarchical controller that learns three expert controllers for robot states, stand, sit, and walk, while manually designing transition controllers between states, “Section III, Subsection C, “In our experiments, it becomes more difficult to obtain an accurate motion mapping when the human motions for multiple tasks are close to each other. To address this issue, we propose to use a hierarchical learning approach that manages a set of expert networks.[79, 74, 52, 18] We first learn three different motion retargeting networks for three robot states, stand, walk, and sit (Figure 4). Each network work can handle multiple tasks, such as manipulation-at-stand or tilting-at stand”, and Figure 4, in particular “Stand” network), (ii) a walking configuration (See at least again above and “Walk” network of Figure 4), (iii) a running configuration, (iv) a crouching configuration (See at least again above and “Sit” network of Figure 4. See also Section V, Subsection D, “Our motion includes multiple tasks, tilting, manipulation, and locomotion, over three different robot states, stand, sit, and walk, which yields different combinations such as tilting at-stand or manipulation-at-sit”. Examiner notes that “crouch” or “crouching” may be a subjective term, particularly in light of any two-legged and four-legged body. However, Applicant appears to provide in [0073] that a pose/configuration which “bring the torso portion closer to (and potentially into contact with) a surface of a floor on which the robot stands” may be considered a crouch or crouching pose/configuration. Thus the “sit” pose, which still allows for “tilting, manipulation, and locomotion” appears to clearly read on “crouching”), (v) a jumping, and (vi) a body roll configuration.
Regarding Claim 28, the combination of Kim and Cashman teaches:
The method of claim 26,
Kim does not explicitly teach, but Cashman teaches:
wherein the machine learning model is trained to learn to map an input comprising a set of joint positions for the first body to a set of joint positions for the second body having a same label (See again at least [0014], [0063], and [0066] for the same reasons as with Claim 26).
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.
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Bashkirov et al. (US 20220148247 A1) which was cited for the reasons found in the rejections found in the first Office Action.
Rubin et al. (US 20250278138 A1) which teaches animation to robot retargeting from user input data using machine learning.
Lu et al. (W. Lu, Y. Liu, J. Sun and L. Sun, "A Motion Retargeting Method for Topologically Different Characters," 2009 Sixth International Conference on Computer Graphics, Imaging and Visualization, Tianjin, China, 2009, pp. 96-100) which teaches motion retargeting between different skeletons, including between bipedal and quadrupedal skeletons in both directions.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MATTHEW C GAMMON whose telephone number is (571)272-4919. The examiner can normally be reached M - F 10:00 - 6:00.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, ADAM MOTT can be reached on (571) 270-5376. 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.
/MATTHEW C GAMMON/Examiner, Art Unit 3657
/ADAM R MOTT/Supervisory Patent Examiner, Art Unit 3657