DETAILED ACTION
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 31 March 2026 has been entered.
Comments
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA . In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
Claims 1-5 and 7-20 are pending and examined in the instant Office action.
The claims are subject matter eligible because the core limitations recite use of neural networks, which is a limitation too complex to be conducted in the human mind. Even assuming (en arguendo) that the claims recite judicial exceptions, the claims recite the practical application of more computationally efficiently analysis of anatomical tree structure data than conventional techniques with analogous objectives.
Claim Interpretation - Means plus function
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The following is a quotation of pre-AIA 35 U.S.C. 112, sixth paragraph:
An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The following comments are necessitated by amendment:
The claims in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is invoked.
This application includes one or more claim limitations that do not use the word “means,” but are nonetheless being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the claim limitation(s) uses a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier. Such claim limitation(s) is/are: “a recurrent neural network (RNN) unit” in each of claims 1, 11, and 19.
Because this/these claim limitation(s) is/are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it/they is/are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof.
If applicant does not intend to have this/these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitation(s) recite(s) sufficient structure to perform the claimed function so as to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph.
Paragraph 30 of the specification teaches that an LSTM or GRU can comprise “a recurrent neural network (RNN) unit”.
Claim Rejections - 35 USC § 112(b) - Indefiniteness
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.
The following rejection is necessitated by amendment:
Claims 1-5 and 7-20 are 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.
Each of independent claims 1, 11, and 19 has been amended to recite that each node of the neural network comprises a recurrent neural network (RNN) that is connected to one or more of the encoders. In this limitation, it is unclear as to whether each node maps to its own distinct RNN (1:1 mapping), or whether there is a different type of mapping (e.g. all of the nodes map to one RNN). For the purpose of examination, the claim is interpreted such that all of the nodes can map to a single RNN.
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.
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.
The following rejection is reiterated for claims 7 and 16 and is necessitated by amendment for claims 1-2, 4-5, 8-12, 14-15, and 17-20:
35 U.S.C. 103 Rejection #1:
Claims 1-2, 4-5, 7-12, and 14-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vunjak-Novakovic et al. [WO 2018/156460 A1] in view of Vion-Dury [US PGPUB 2013/0151565 A1] in view of Sundaram [US Patent 9,972,339 B1].
Claim 1 is drawn to a method for an anatomical tree structure analysis. The method comprises receiving model inputs related to an anatomical tree structure for a set of positions in the anatomical tree structure. The anatomical tree structure includes at least one bifurcation point and a plurality of branches splitting from the at least one bifurcation point. The method comprises applying a learning network to the model inputs. The learning network comprises a set of encoders and a neural network modeling the anatomical tree structure. Each encoder is external to the neural network and extracts features from one of the model inputs at a corresponding position in the set of positions in the anatomical tree structure. The neural network has a plurality of nodes constructed according to the set of positions in the anatomical tree structure and each node of the neural network comprises a recurrent neural network (RNN) unit that is connected to one or more of the encoders and is configured to receive and process the extracted features from one or more of the encoders. The method comprises providing an output of the tree structured network as an analysis result of the anatomical tree structure analysis.
Claim 11 is drawn to similar subject matter as claim 1, except claim 11 is drawn to a system.
Claim 19 is drawn to similar subject matter as claim 1, except claim 19 is drawn to a non-transitory computer-readable medium.
Claims 2, 12, and 20 are further limiting wherein the anatomical tree structure comprises an airway.
Claims 5 and 15 are further limiting wherein the encoders comprise CNNs.
The document of Vunjak-Novakovic et al. studies a method and apparatus for computer-vision guided targeted delivery of small liquid volumes into selected lung regions [title]. Figure 1 of Vunjak-Novakovic et al. illustrates the anatomical tree-like lung structure wherein the encoders external to the human and the lungs extracts features from at least one of the positions of the anatomical tree like lung structure. Paragraph 138 of Vunjak-Novakovic et al. teaches application of CNNs to input training sets of airway tree bifurcation data to predict the bifurcations of a new tree structure of a lung that has yet to be classified. It is evident that a lung has a plurality of bifurcation points with a plurality of branches splitting from at least one bifurcation point. Each bifurcation point in a lung is interpreted to be a node. Figure 24 of Vunjak-Novakovic et al. teaches output of the tree structured network as an analysis result of the anatomical tree structure analysis.
Vunjak-Novakovic et al. does not teach that the plurality of nodes is constructed according to the set of positions in the anatomical tree structure. Vunjak-Novakovic et al. does not teach that each node of the neural network comprises an RNN.
The document of Vion-Dury studies arithmetic node encoding for tree structures [title]. Figure 1 of Vion-Dury illustrates an upside-down tree structure of nodes. Figure 1, the abstract, and paragraphs 6-9 of Vion-Dury teaches that each node encodes data (i.e. the nodes also act as encoders of data).
Vunjak-Novakovic et al. and Vion-Dury do not teach that each node of the neural network comprises an RNN.
The document of Sundaram studies neural network based beam selection [title]. Column 18, line 41 to column 19, line 13 of Sundaram teaches RNN, LSTM, GRU, bi-directionality between nodes, and single directionality between nodes.
With regard to claims 4, 10, and 14, paragraph 138 of Vunjak-Novakovic et al. teaches receiving images of the lung as acquired by an image acquisition device, deriving model inputs at the set of positions in the anatomical tree structure, and using the thousands of images to jointly train a new model of the lung.
Claim 16 is further limiting comprising using RNN. Claim 7 is further limiting wherein the RNN units are selected from LSTM and GRU.
The document of Sundaram studies neural network based beam selection [title]. Column 18, line 41 to column 19, line 13 of Sundaram teaches RNN, LSTM, GRU, bi-directionality between nodes, and single directionality between nodes.
With regard to claims 8-9 and 17-18, Figure 24 of Vunjak-Novakovic et al. illustrates image patches and anatomical structure labeling.
It would have been obvious to someone of ordinary skill in the art at the time of the effective filing date of the instant application to modify the machine learning of anatomical tree structures of Vunjak-Novakovic et al. by use of the nodes encoding data of Vion-Dury wherein the motivation would have been that Vion-Dury gives additional mathematical tools to assist with the machine learning of Vunjak-Novakovic et al. [paragraph 6-9 of Vion-Dury]. There would have been a reasonable expectation of success in combining Vunjak-Novakovic et al. with Vion-Dury because both studies are analogously applicable to studying biological applications of trees of data structure.
It would have been obvious to someone of ordinary skill in the art at the time of the effective filing date of the instant application to modify the machine learning of anatomical tree structures of Vunjak-Novakovic et al. and the nodes encoding data of Vion-Dury by use of the RNN, LSTM, GRU, bi-directionality between nodes, and single directionality between nodes of Sundaram wherein the motivation would have been that Sundaram teaches additional tools and features that facilitates analysis of a graph [column 18, line 41 to column 19, line 13 of Sundaram]. There would have been a reasonable expectation of success in combining Vunjak-Novakovic et al., Vion-Dury, and Sundaram because all three studies are analogously applicable to using analysis of graphs and/or tree-like structures.
Response to arguments:
Applicant's arguments filed 31 March 2026 have been fully considered but they are not persuasive.
Applicant generally asserts that the CNN does not model the anatomical tree structure in Vunjak-Novakovic et al., but lacks any supporting reasoning.
Applicant argues that the machine learning of Vunjak-Novakovic et al. lacks the external encoders as recited in the claim. This limitation relies of the teachings of Vion-Dury.
Applicant argues that the amendments to the independent claims overcome the prior art rejections. This argument is not persuasive because, in view of the indefiniteness rejection above, Sundaram teaches the association of nodes/encoders with an RNN.
Applicant argues that Sundaram is non-analogous art with no relevance to anatomical tree structures. This argument is not persuasive because the teaching of the association of nodes/encoders to an RNN of Sundaram is robust and generally applicable to the problem of associating nodes with neural networks as demonstrated in Vunjak-Novakovic et al. and Vion Dury.
The following rejection is reiterated:
35 U.S.C. 103 Rejection #2:
Claims 3 and 13 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vunjak-Novakovic et al. in view of Vion-Dury in view of Sundaram as applied to claims 1-2, 4-5, 7-12, and 14-20 above, in further view of Tsukahara et al. [US PGPUB 2018/0255284 A1].
Claims 3 and 13 are further limiting wherein the set of positions include the at least one bifurcation point and at least one point in each branch.
The documents of Vunjak-Novakovic et al., Vion-Dury, and Sundaram study using machine learning to analyze the anatomical tree structure of the lung, as discussed above.
Vunjak-Novakovic et al., Vion-Dury, and Sundaram do not teach that the set of positions include the at least one bifurcation point and at least one point in each branch.
The document of Tsukahara et al. studies an information processing apparatus, information processing method, and program [title]. Figure 4 of Tsukahara et al. illustrates a tree structure wherein the set of positions include the at least one bifurcation point and at least one point in each branch.
It would have been obvious to someone of ordinary skill in the art at the time of the effective filing date of the instant application to modify the machine learning of anatomical tree structures of Vunjak-Novakovic et al., the nodes encoding data of Vion-Dury, and the RNN of Sundaram by use of the tree structure of Tsukahara et al. because it is obvious to combine known elements in the prior art to yield a predictable result. In this instance, the tree structure of Tsukahara et al. is an alternative to the lungs of Vunjak-Novakovic et al. There would have been a reasonable expectation of success in combining Vunjak-Novakovic et al., Vion-Dury, Sundaram, and Tsukahara et al. because all four studies are analogously applicable to analysis of tree structures.
Response to arguments:
Applicant's arguments filed 31 March 2026 have been fully considered but they are not persuasive.
Applicant argues that Tsukahara et al. does not overcome the alleged deficiencies of the initial obviousness prior art rejection. This argument is not persuasive because the initial obviousness prior art rejection is not deficient.
The following rejection is reiterated:
35 U.S.C. 103 Rejection #3:
Claims 1-5 and 7-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Wang et al. [CN 107977709 A] in view of Vion-Dury [US PGPUB 2013/0151565 A1]. An English machine translation of Wang et al. is cited in the instant Office action.
Claim 1 is drawn to a method for an anatomical tree structure analysis. The method comprises receiving model inputs related to an anatomical tree structure for a set of positions in the anatomical tree structure. The anatomical tree structure includes at least one bifurcation point and a plurality of branches splitting from the at least one bifurcation point. The method comprises applying a learning network to the model inputs. The learning network comprises a set of encoders and a neural network modeling the anatomical tree structure. Each encoder is external to the neural network and extracts features from one of the model inputs at a corresponding position in the set of positions in the anatomical tree structure. The neural network has a plurality of nodes constructed according to the set of positions in the anatomical tree structure and each node of the neural network comprises a recurrent neural network (RNN) unit that is connected to one or more of the encoders and is configured to receive and process the extracted features from one or more of the encoders. The method comprises providing an output of the tree structured network as an analysis result of the anatomical tree structure analysis.
Claim 11 is drawn to similar subject matter as claim 1, except claim 11 is drawn to a system.
Claim 19 is drawn to similar subject matter as claim 1, except claim 19 is drawn to a non-transitory computer-readable medium.
Claims 2, 12, and 20 are further limiting wherein the anatomical tree structure comprises an airway.
Claims 3 and 13 are further limiting wherein each branch has a bifurcation point.
Claims 5 and 15 are further limiting wherein the encoders comprise CNNs.
The document of Wang et al. studies a method and system to predict the deep learning model and system of the flow characteristic on vascular tree blood flow paths [title]. The abstract of Wang et al. teaches a neural network performed at nodes on the blood flow paths in order to extract features and images of the blood flow paths. Figures 1 and 2 of Wang et al. illustrate the anatomical tree-like lung structure wherein the encoders external to the anatomic tree and extract features from at least one of the positions of the anatomical tree like lung structure. Figures 1 and 2 of Wang et al. illustrate the neural network set-up on the blood flow path wherein the blood flow path has a bifurcation point (with a corresponding node/neural network). Figures 1 and 2 of Wang et al. teaches nodes and neural networks associates with points on the tree-like bifurcated structure. The description of Figure 2 on page 4 of Wang et al. teaches use of CNNs. The abstract of Wang et al. teaches use of RNNs. The description of Figure 2 on page 4 of Wang et al. teaches that the RNNs are selected from LSTM and GRU.
Wang et al. does not teach encoders associated with the nodes.
The document of Vion-Dury studies arithmetic node encoding for tree structures [title]. Figure 1 of Vion-Dury illustrates an upside-down tree structure of nodes. Figure 1, the abstract, and paragraphs 6-9 of Vion-Dury teaches that each node encodes data (i.e. the nodes also act as encoders of data).
With regard to claims 4 and 14, the abstract of Wang et al. teaches acquiring an image feature at each point. Figures 1 and 2 of Wang et al. teach model inputs at a set of positions in the anatomical tree structure.
With regard to claims 7 and 16, the abstract of Wang et al. teaches use of RNNs. The description of Figure 2 on page 4 of Wang et al. teaches that the RNNs are selected from LSTM and GRU.
With regard to claims 8-10 and 17-18, page 4 of Wang et al. teaches image features as inputs. Figures 1 and 2 of Wang et al. teach anatomical structure labeling. Figures 1 and 2 of Wang et al. teach joint training of neural networks.
It would have been obvious to someone of ordinary skill in the art at the time of the effective filing date of the instant application to modify the machine learning of anatomical tree structures of Wang et al. by use of the nodes encoding data of Vion-Dury wherein the motivation would have been that Vion-Dury gives additional mathematical tools to assist with the machine learning of Wang et al. [paragraph 6-9 of Vion-Dury]. There would have been a reasonable expectation of success in combining Wang et al. with Vion-Dury because both studies are analogously applicable to studying biological applications of trees of data structure.
Response to arguments:
Applicant's arguments filed 31 March 2026 have been fully considered but they are not persuasive.
Applicant argues that the 35 U.S.C. 102(b)(1)(A) exception applies to Wang et al. because of overlapping inventive entities. This argument is not persuasive because the inventive entity of Wang et al. is not a subset of the inventive entity of the instant application. In other words, Yuwei Li and Dan Wu are inventors on the prior art of Wang et al., but are not inventors on the instant application. Therefore, the 35 U.S.C. 102(b)(1)(A) exception does not automatically apply.
Applicant argues that Figure 2 of Wang et al. does not teach a plurality of nodes at set positions across an anatomical tree structure. This argument is not persuasive because Figure 2 of Wang et al. maps nodes and machine learning to different locations across an anatomical tree structure wherein the tree structure has a bifurcation point. The claims do not require that nodes map to every branch of the anatomical tree structure, but instead maps according to a set of positions in the anatomical tree structure.
Applicant argues that Figure 2 of Wang et al, does not teach the amended limitation of the independent claims. In view of the indefiniteness rejection above, the description of Figure 2 on page 4 of Wang et al. are broadly construed to teach that the RNNs are selected from LSTM and GRU and are associated with the encoders and nodes.
Related Prior Art
The prior art of Kalchbrenner et al. [arXiv:1404.2188v1; 8 April 2014, 11 pages; on IDS] teaches applying CNNs to tree structures representing sentences and not blood vessels or airways. However, instead of anatomical modeling, Kalchbrenner et al. uses advanced machine learning to model the linguistics of sentences.
E-mail Communications Authorization
Per updated USPTO Internet usage policies, Applicant and/or applicant’s representative is encouraged to authorize the USPTO examiner to discuss any subject matter concerning the above application via Internet e-mail communications. See MPEP 502.03. To approve such communications, Applicant must provide written authorization for e-mail communication by submitting the following statement via EFS-Web (using PTO/SB/439) or Central Fax (571-273-8300):
Recognizing that Internet communications are not secure, I hereby authorize the USPTO to communicate with the undersigned and practitioners in accordance with 37 CFR 1.33 and 37 CFR 1.34 concerning any subject matter of this application by video conferencing, instant messaging, or electronic mail. I understand that a copy of these communications will be made of record in the application file.
Written authorizations submitted to the Examiner via e-mail are NOT proper. Written authorizations must be submitted via EFS-Web (using PTO/SB/439) or Central Fax (571-273-8300). A paper copy of e-mail correspondence will be placed in the patent application when appropriate. E-mails from the USPTO are for the sole use of the intended recipient, and may contain information subject to the confidentiality requirement set forth in 35 USC § 122. See also MPEP 502.03.
Conclusion
No claim is allowed.
Any inquiry concerning this communication or earlier communications from the Examiner should be directed to Russell Negin, whose telephone number is (571) 272-1083. This Examiner can normally be reached from Monday through Thursday from 8 am to 3 pm and variable hours on Fridays.
If attempts to reach the Examiner by telephone are unsuccessful, the Examiner’s Supervisor, Larry Riggs, Supervisory Patent Examiner, can be reached at (571) 270-3062.
/RUSSELL S NEGIN/ Primary Examiner, Art Unit 1686 11 July 2026
Prosecution Insights