The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
DETAILED ACTION
Double Patenting
The nonstatutory double patenting rejection is based on a judicially created doctrine grounded in public policy (a policy reflected in the statute) so as to prevent the unjustified or improper timewise extension of the “right to exclude” granted by a patent and to prevent possible harassment by multiple assignees. A nonstatutory double patenting rejection is appropriate where the conflicting claims are not identical, but at least one examined application claim is not patentably distinct from the reference claim(s) because the examined application claim is either anticipated by, or would have been obvious over, the reference claim(s). See, e.g., In re Berg, 140 F.3d 1428, 46 USPQ2d 1226 (Fed. Cir. 1998); In re Goodman, 11 F.3d 1046, 29 USPQ2d 2010 (Fed. Cir. 1993); In re Longi, 759 F.2d 887, 225 USPQ 645 (Fed. Cir. 1985); In re Van Ornum, 686 F.2d 937, 214 USPQ 761 (CCPA 1982); In re Vogel, 422 F.2d 438, 164 USPQ 619 (CCPA 1970); In re Thorington, 418 F.2d 528, 163 USPQ 644 (CCPA 1969).
A timely filed terminal disclaimer in compliance with 37 CFR 1.321(c) or 1.321(d) may be used to overcome an actual or provisional rejection based on nonstatutory double patenting provided the reference application or patent either is shown to be commonly owned with the examined application, or claims an invention made as a result of activities undertaken within the scope of a joint research agreement. See MPEP § 717.02 for applications subject to examination under the first inventor to file provisions of the AIA as explained in MPEP § 2159. See MPEP §§ 706.02(l)(1) - 706.02(l)(3) for applications not subject to examination under the first inventor to file provisions of the AIA . A terminal disclaimer must be signed in compliance with 37 CFR 1.321(b).
The USPTO Internet website contains terminal disclaimer forms which may be used. Please visit www.uspto.gov/patent/patents-forms. The filing date of the application in which the form is filed determines what form (e.g., PTO/SB/25, PTO/SB/26, PTO/AIA /25, or PTO/AIA /26) should be used. A web-based eTerminal Disclaimer may be filled out completely online using web-screens. An eTerminal Disclaimer that meets all requirements is auto-processed and approved immediately upon submission. For more information about eTerminal Disclaimers, refer to www.uspto.gov/patents/process/file/efs/guidance/eTD-info-I.jsp.
Claims 1-37 are rejected on the ground of nonstatutory double patenting as being unpatentable over claims 1-37 of U.S. Patent No. US12266073B2.
Although the conflicting claims are not identical, they are not patentably distinct from each other because the patent US12266073B2 contains substantially all the limitations of the instant application claims.
Application No. 19/065888
Patent No. US12266073B2
1. A garment rendering system, comprising a virtual garment generation unit for receiving garment related data and for generating in response to the garment related data virtual garment data, wherein the garment related data includes garment image data and garment textual data, an avatar generation unit for receiving user data and for generating, in response to the user data, user avatar data representative of an avatar of the user, wherein the user data includes at least one of user image data and user biometric data,a simulation engine for receiving the user avatar data and the virtual garment data and for generating in response thereto simulated garment data representative of a simulated garment draped on the avatar, and a rendering unit for rendering the simulated garment data received from the simulation engine and for generating a rendered virtual image, wherein the simulation engine includes a tailor unit for receiving the virtual garment data and encoding the virtual garment data with the garment image data to form encoded garment data, a graph network configured for receiving the encoded garment data and for generating garment related training data, and a one-shot network configured for receiving the garment related training data in order to train the one-shot network, and wherein the one-shot network is configured for receiving the virtual garment data and the user avatar data and for generating based thereon the simulated garment data.
1. A garment rendering system, comprising
a virtual garment generation unit for receiving garment related data and for generating in response to the garment related data virtual garment data, wherein the garment related data includes garment image data and garment textual data,
an avatar generation unit for receiving user data and for generating, in response to the user data, user avatar data representative of an avatar of the user, wherein the user data includes at least one of user image data and user biometric data,
a simulation engine for receiving the user avatar data and the virtual garment data and for generating in response thereto simulated garment data representative of a simulated garment draped on the avatar, and
a rendering unit for rendering the simulated garment data received from the simulation engine and for generating a rendered virtual image,
wherein the simulation engine includes
a tailor unit for receiving the virtual garment data and encoding the virtual garment data with the garment image data to form encoded garment data,
a graph network configured for receiving the encoded garment data and for generating garment related training data, and
a one-shot network configured for receiving the garment related training data in order to train the one-shot network, and wherein the one-shot network is configured for receiving the virtual garment data and the user avatar data and for generating based thereon the simulated garment data.
15. A method for generating a rendered virtual image of a garment on an avatar of a user, comprising generating, in response to receiving garment related data, virtual garment data with a virtual garment generation unit, wherein the garment related data includes garment image data and garment textual data,generating with an avatar generation unit, in response to receiving user data, user avatar data that is representative of the avatar of the user, wherein the user data includes at least one of user image data and user biometric data,generating with a simulation engine, in response to receiving the user avatar data and the virtual garment data, simulated garment data representative of a simulated garment draped on the avatar, and rendering the simulated garment data with a rendering unit and generating the rendered virtual image, wherein generating with the simulation engine includes encoding with a tailor unit the virtual garment data with the garment image data to form encoded garment data,providing a graph network configured for receiving the encoded garment data and for generating garment related training data, andproviding a one-shot network configured for receiving the garment related training data in order to train the one-shot network, and wherein the one-shot network is configured for receiving the virtual garment data and the user avatar data and for generating based thereon the simulated garment data.
15. A method for generating a rendered virtual image of a garment on an avatar of a user, comprising
generating, in response to receiving garment related data, virtual garment data with a virtual garment generation unit, wherein the garment related data includes garment image data and garment textual data,
generating with an avatar generation unit, in response to receiving user data, user avatar data that is representative of the avatar of the user, wherein the user data includes at least one of user image data and user biometric data,
generating with a simulation engine, in response to receiving the user avatar data and the virtual garment data, simulated garment data representative of a simulated garment draped on the avatar, and
rendering the simulated garment data with a rendering unit and generating the rendered virtual image,
wherein generating with the simulation engine includes
encoding with a tailor unit the virtual garment data with the garment image data to form encoded garment data,
providing a graph network configured for receiving the encoded garment data and for generating garment related training data, and
providing a one-shot network configured for receiving the garment related training data in order to train the one-shot network, and wherein the one-shot network is configured for receiving the virtual garment data and the user avatar data and for generating based thereon the simulated garment data.
28. A non-transitory, computer readable medium comprising computer program instructions tangibly stored on the computer readable medium, wherein the computer program instructions are executable by at least one computer processor to perform a method, the method comprising: generating, in response to receiving garment related data, virtual garment data with a virtual garment generation unit, wherein the garment related data includes garment image data and garment textual data, generating with an avatar generation unit, in response to receiving user data, user avatar data that is representative of the avatar of the user, wherein the user data includes at least one of user image data and user biometric data, generating with a simulation engine, in response to receiving the user avatar data and the virtual garment data, simulated garment data representative of a simulated garment draped on the avatar, and rendering the simulated garment data with a rendering unit and generating the rendered virtual image wherein generating with the simulation engine includes encoding with a tailor unit the virtual garment data with the garment image data to form encoded garment data, providing a graph network configured for receiving the encoded garment data and for generating garment related training data, and providing a one-shot network configured for receiving the garment related training data in order to train the one-shot network, and wherein the one-shot network is configured for receiving the virtual garment data and the user avatar data and for generating based thereon the simulated garment data.
28. A non-transitory, computer readable medium comprising computer program instructions tangibly stored on the computer readable medium, wherein the computer program instructions are executable by at least one computer processor to perform a method, the method comprising:
generating, in response to receiving garment related data, virtual garment data with a virtual garment generation unit, wherein the garment related data includes garment image data and garment textual data,
generating with an avatar generation unit, in response to receiving user data, user avatar data that is representative of the avatar of the user, wherein the user data includes at least one of user image data and user biometric data,
generating with a simulation engine, in response to receiving the user avatar data and the virtual garment data, simulated garment data representative of a simulated garment draped on the avatar, and
rendering the simulated garment data with a rendering unit and generating the rendered virtual image
wherein generating with the simulation engine includes
encoding with a tailor unit the virtual garment data with the garment image data to form encoded garment data,
providing a graph network configured for receiving the encoded garment data and for generating garment related training data, and
providing a one-shot network configured for receiving the garment related training data in order to train the one-shot network, and wherein the one-shot network is configured for receiving the virtual garment data and the user avatar data and for generating based thereon the simulated garment data.
Dependent claims 2-14, 16-27, 29-37 recite similar matter as claim 1-37 of Patent US12266073B2 and are rejected for the same reason.
35 USC § 112 (f)
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.
Use of the word “means” (or “step for”) in a claim with functional language creates a rebuttable presumption that the claim element is to be treated in accordance with 35 U.S.C. § 112(f) (pre-AIA 35 U.S.C. 112, sixth paragraph). The presumption that § 112(f) (pre-AIA § 112, sixth paragraph) is invoked is rebutted when the function is recited with sufficient structure, material, or acts within the claim itself to entirely perform the recited function.
Absence of the word “means” (or “step for”) in a claim creates a rebuttable presumption that the claim element is not to be treated in accordance with 35 U.S.C. § 112(f) (pre-AIA 35 U.S.C. 112, sixth paragraph). The presumption that § 112(f) (pre-AIA § 112, sixth paragraph) is not invoked is rebutted when the claim element recites function but fails to recite sufficiently definite structure, material or acts to perform that function.
Regarding claims 1-14, the claim limitations, " a virtual garment generation unit for", “an avatar generation unit for”, “a simulation engine for” and "a rendering unit for” have been interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because they use a generic placeholder, "unit for" coupled with functional language, "receiving" "generating” and “rendering” without reciting sufficient structure to achieve the function.
Since this claim limitation invokes 35 U.S.C. 112(f) or 35 U.S.C. 112 (pre-AIA ), sixth paragraph, claims 1-14 are interpreted to cover the corresponding structure described in the specification that achieves the claimed function, and equivalents thereof.
A review of the specification shows that the following appears to be the corresponding structure described in the specification for the 35 U.S.C. 112(f) or 35 U.S.C. 112 (pre-AIA ), sixth paragraph limitation: ¶0086 of the specification in patent publication discusses method steps of the invention may be performed by one or more computer processors executing a program tangibly embodied on a computer-readable medium to perform functions of the invention by operating on input and generating output. Suitable processors include, by way of example, both general and special purpose microprocessors.
If applicant wishes to provide further explanation or dispute the examiner’s interpretation of the corresponding structure, applicant must identify the corresponding structure with reference to the specification by page and line number, and to the drawing, if any, by reference characters in response to this Office action.
If applicant does not wish to have the claim limitation treated under 35 U.S.C. 112(f) or 35 U.S.C. 112 (pre-AIA ), sixth paragraph, applicant may amend the claim so that it will clearly not invoke 35 U.S.C. 112(f) or 35 U.S.C. 112 (pre-AIA ), sixth paragraph, or present a sufficient showing that the claim recites sufficient structure, material, or acts for performing the claimed function to preclude application of 35 U.S.C. 112(f) or 35 U.S.C. 112 (pre-AIA ), sixth paragraph.
For more information, see MPEP § 2173 et seq. and Supplementary Examination Guidelines for Determining Compliance with 35 U.S.C. § 112 and for Treatment of Related Issues in Patent Applications, 76 FR 7162, 7167 (Feb. 9, 2011).
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 of this title, 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 set forth in Graham v. John Deere Co., 383 U.S. 1, 148 USPQ 459 (1966), that are applied 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.
Claims 1-37 are rejected under 35 U.S.C. 103 as being unpatentable over Chen et al. (US Pub 2020/0066029 A1), as Chen A in view of Wang, T. Y., Ceylan, D., Popovic, J., & Mitra, N. J. (2018). (Learning a shared shape space for multimodal garment design. arXiv preprint arXiv:1806.11335.) and Chen, T., Ji, B., Ding, T., Fang, B., Wang, G., Zhu, Z., ... & Tu, X. (2021). (Only train once: A one-shot neural network training and pruning framework. Advances in Neural Information Processing Systems, 34, 19637-19651.), as Chen B.
As to claim 1, Chen A discloses a garment rendering system, comprising
a virtual garment generation unit for receiving garment related data and for generating in response to the garment related data virtual garment data, wherein the garment related data includes garment image data and garment textual data (Chen A, ¶0058,” provide a production system for producing a garment via use of an interactive garment model created from one or more two dimensional images. The production system includes a data processing arrangement which executes one or more software products to create an interactive environment for hosting the interactive garment model. The two dimensional images may comprise at least a garment, a model wearing the garment and a background, by way of content. Furthermore, the production system may acquire the one or more two dimensional images of a garment from one or more viewing angles and create a three dimensional model of the garment.” ¶0085, “The one or more two dimensional images may include e-commerce product photos on retailer websites, existing mannequin or model photography from retailers, and images of dressed person(s) obtained from the Internet.” ¶0120, “the garment database may be configured to provide one or more digital garment models comprising the texture, geometry, finite element representations, and other metadata.” ¶0123, “the data processing arrangement generates the digital garment model to comprise one or more of: garment geometry information, a 2D silhouette, a 3D mesh model, garment appearance information, cut-out garment texture sprites, texture maps for a 3D mesh, alpha maps, garment metadata information, landmark points, control splines, garment style attributes, garment colour attributes, garment texture attributes, garment shape attributes, garment fit styles attributes, garment vertical drape attributes, a garment cuff, a garment hem height, a garment waist line position, a garment material stretch, garment strain attributes.”),
an avatar generation unit for receiving user data and for generating, in response to the user data, user avatar data representative of an avatar of the user, wherein the user data includes at least one of user image data and user biometric data (Chen A, ¶0118-0119, “the data processing arrangement receives inputs from the customer, wherein the inputs include one or more consumer body metrics, and the second subsystem is configured to generate automatically a 3D body avatar representing the consumer's body shape.” ¶0181, “The body pose matching and pose changing module 114 may be configured to detect the underlying body pose in an input image and adjust the pose of a 3D virtual body model to match that in the input image.”),
a simulation engine for receiving the user avatar data and the virtual garment data and for generating in response thereto simulated garment data representative of a simulated garment draped on the avatar (Chen A, ¶0120, “the second subsystem selects automatically one or more digital garment models from a garment database which match a consumer's garment choice input to the production system. Accordingly, in some embodiments, the garment database may be configured to provide one or more digital garment models comprising the texture, geometry, finite element representations, and other metadata.” ¶0121, “the second subsystem simulates automatically a drape and a fit of the digital garment model onto the 3D body avatar. The second subsystem may simulate the drape and the fit of the digital garment model onto the 3D body avatar using finite-element analysis. In some embodiments, the garment physics simulation module may be configured to perform finite-element analysis of the digital garment model(s) on the generated 3D body avatar of the customers. At the runtime, the garment physics simulation module models how the garment model will drape and fit against, or on, a customer body shape.”), and
a rendering unit for rendering the simulated garment data received from the simulation engine and for generating a rendered virtual image (Chen A, ¶0122, “the second subsystem composites automatically multiple individual garment models in a correct layering order and then displays the 3D body avatar in an outfit of one or more selected garments. In some embodiments, the outfitting and rendering module may be configured to composite multiple individual garment models in the correct layering order and then display the 3D body avatar of the customer in an outfit of one or more selected garments. The output of the second subsystem is a dressed body avatar(s) (in 2D/3D) visualization with the target garment.” ¶0145, “rendering a dressed avatar of the customer for final visualization”),
wherein the simulation engine includes a tailor unit for receiving the virtual garment data and encoding the virtual garment data with the garment image data to form encoded garment data (Chen A, ¶0105, “The first stage of the two-stage algorithm includes using performing deep neural network coarse segmentation on a downsized image (such as 512×512 size image). As a method of implementation, the deep neural network models (e.g. SegNet [e.g. V. Badrinarayanan, A. Kendall, and R. Cipolla: “Segnet: A deep convolutional encoder-decoder architecture for image segmentation”. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017], DeepLab [e.g. L. Chen, G. Papandreou, L. Kokkinos, K. Murphy, and A. L. Yuille: “DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs”, arXiv preprint, 2016] may be trained using the product photos in Red Green Blue (RGB) or Red Green Blue Alpha (RGBA) formats as input and corresponding manually-labelled semantic label image as outputs.”),
Chen A does not disclose a graph network configured for receiving the encoded garment data and for generating garment related training data, and
a one-shot network configured for receiving the garment related training data in order to train the one-shot network, and wherein the one-shot network is configured for receiving the virtual garment data and the user avatar data and for generating based thereon the simulated garment data.
Wang teaches a graph network configured for receiving the encoded garment data and for generating garment related training data (Wang, Page 2, “We achieve this by jointly training multiple encoder-decoder networks that each specializes at linking pairs of representations (e.g., recovering garment and body parameters from a sketch or recovering the on-body shapes of garments from the parameters) while operating at a common embedding. To train these network, we create a large scale synthetic dataset. Specically, we first define a set of parameterized garment types (shirt, skirt, and kimono) and generate different garments by sampling this representation. Then, we simulate each garment on a set of 3D body shapes and poses sampled from a deformable body model. Finally, for each of these simulated examples, we generate 2D sketches using non-photorealistic rendering. Thus, each examplar triplet in our dataset includes (i) a 2D sketch, (ii) garment and body parameters, and (iii) the resultant draped 3D shape of the garment. Subsequently, by jointly training multiple encoder-decoder networks via a novel multimodal loss function, we learn a common embedding that can
be queried using any of the different modalities.”) and
a (Wang, Page 5, “(i) sketch descriptor to latent space (FS2L = S → L), (ii) parameter space to latent space (FP2L = P → L), (iii) latent space to parameter space (FL2P = L → P), and (iv) latent space to the draped garment shape (FL2M = L → M). We learn these mappings by jointly training four encoder-decoder neural networks (i.e., sketch-to-parameter, sketch-to-3D garment shape, parameter to-3D garment shape, and parameter-to-parameter) that share a common embedding space (see Figure 6). We describe the network architecture with more details in section 4.”).
Chen A and Wang are considered to be analogous art because all pertain deep neural network. It would have been obvious before the effective filing date of the claimed invention to have modified Chen A with the features of “a graph network configured for receiving the encoded garment data and for generating garment related training data, and a Wang, page 5).
Chen A does not disclose a one-shot network.
Chen B teaches a one-shot network (Chen B, abstract, “compresses DNNs into slimmer architectures with competitive performances and significant FLOPs reductions by Only-Train-Once (OTO).”).
Chen A, Wang and Chen B are considered to be analogous art because all pertain to 3-D garment modeling. It would have been obvious before the effective filing date of the claimed invention to have modified Chen A with the features of “a one-shot network.” as taught by Chen B. The suggestion/motivation would have been in order to dramatically simplifies the complex multi-stage training pipelines of the existing pruning approaches, fits various architectures and applications, and hence is generic and efficient (Chen B, page 2).
As to claim 2, claim 1 is incorporated and the combination of Chen A, Wang and Chen B discloses the virtual garment generation unit comprises a pattern determination unit that is configured to receive the garment image data and the garment textual data and to generate two-dimensional garment pattern data based on the garment image and the garment textual data (Chen A, ¶0091, “the data processing arrangement automatically identifies and recreates any occluded sections of the garment within the two dimensional image to produce a non-occluded garment. Accordingly, in some embodiments, the garment texture in-painting module may be used to semi-automatically complete any occluded garment texture (or occluded sections) in a product photo using an interactive UI. The occluded textures may result due to occluding hair or arms of the model. Accordingly, the garment texture in-painting module may fill up the occluded texture either by cloning the texture samples or by interpolating the colours from neighbouring unoccluded areas. A goal of such a method is to generate an overall visually plausible and coherent texture of the whole garment.” ¶0092, “The patch based algorithms may be used for handling repetitive patterns or random patterns.” ¶0093, “The diffusion-based algorithms may be used for handling uniform-coloured patterns with gradual shading variations.” ¶0151, “The garment texture in-painting module 104 may fill up the occluded texture either by cloning the texture samples or by interpolating the colours from neighbouring unoccluded areas. The goal is to generate an overall visually plausible and coherent texture of the whole garment. Accordingly, two streams of image processing based algorithms may be implemented for handling different types of garment patterns.”),
a mesher unit for receiving and processing the two dimensional garment pattern data and for generating, based thereon, a plurality of three-dimensional garment meshes forming at least part of the virtual garment data (Chen A, ¶0039, “generating the complete 3D garment model to comprise one or more of: garment geometry information, a 2D silhouette, a 3D mesh model, garment appearance information, cut-out garment texture sprites, texture maps for a 3D mesh, alpha maps, garment metadata information, landmark points, control splines, garment style attributes, garment colour attributes, garment texture attributes, garment shape attributes, garment fit styles attributes, garment vertical drape attributes, a garment cuff, a garment hem height, a garment waist line position, a garment material stretch, garment strain attributes.” ¶0189, “the digital garment model 128 may include garment geometry data 130, garment appearance data 132 and garment metadata 134. The garment geometry data 130 may include, but is not limited to, 2D silhouette(s) and 3D mesh model(s). The garment appearance data 132 may include, but is not limited, to cut-out garment texture sprites, texture maps for 3D mesh generated by back-projecting and stitching the 2D texture from each view, and alpha maps capturing the translucency of the garment. The garment metadata 134 may include landmark points and control splines, which are useful for instructing physics simulation and layer composition in an outfit. Further, the garment metadata 134 may include garment attributes of style, colour, texture, shape, fit styles, vertical drapes (e.g. cuff, hem height, waist line position) material stretch and strain, etc.”), and
a material determination unit configured to determine a type of material associated with the garment based on the garment image data and the garment textual data (Chen A, ¶0133, “material stretch attributes, material strain attributes.” ¶0191, “The labelled garment image dataset 122 is a structured dataset containing a volume of garment image data and corresponding manually-generated semantic label map (indicating the regions of garment, skin, hair, background, etc.), attributes (including garment category, pattern, colour, material properties, camera view, etc.), control point/spline annotations (including garment landmark, body joints, skeleton layout, etc). The labelled garment image dataset 122 may be used as training data for learning machine learning models for garment segmentation, automatic landmark/control spline generation, body pose detection, camera view classification, etc.” Wang, Page 1, “namely 2D sketches for initial design, 2D sewing patterns and material selection for parameterized modeling, and 3D garment shape to model the interaction between garments and subject bodies producing on-body garments” Page 5, “material properties, i.e., stretch, blend, and shear stiffness (Figure 5).” Page 6, “3 material parameters and 10 body shape parameters are consistent across the different garment types”).
As to claim 3, claim 2 is incorporated and the combination of Chen A, Wang and Chen B discloses the mesher unit employs one or more meshing techniques for subdividing the two-dimensional garment pattern data into the plurality of three-dimensional garment meshes (Chen A, ¶0039, “texture maps for a 3D mesh.” ¶0105, “the garment cut-out and layer dividing module may be used to split further the garment texture into one or more layers corresponding to one or more body parts.” ¶0106, “the garment cut-out and layer dividing module” ¶0114, “apply a spatial transformation on the cut-out garment texture sprites so that it may be laid on the proper positions on the standard body avatar used for garment digitisation.” ¶0163, “segmented garment textures” ¶0189.).
As to claim 4, claim 3 is incorporated and the combination of Chen A, Wang and Chen B discloses the garment related data is indicative at least one garment, and wherein the pattern determination unit comprises a classification unit for classifying the garment into one of a plurality of different categories indicative of different types of garments and for generating in response garment classification data (Chen A, ¶0108, “The camera view classification module may be implemented using a machine learning classifier (e.g. random forests [e.g. J. Shotton, M. Johnson, R. Cipolla: “Semantic Texton Forests for Image Categorization and Segmentation” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Jun. 1, 2008], or standard deep neural networks for classification [e.g. C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V. Vanhoucke and A. Rabinovich: “Going deeper with convolutions”. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 1-9, 2015], [e.g. A. Krizhevsky, I. Sutskever, and G. E. Hinton: “Imagenet classification with deep convolutional neural networks”. NIPS, 1(2):4, 2012] which may be trained on a number of labelled model image data of different camera views.”),
an object detection unit configured to receive the garment classification data and for detecting and extracting at least one garment segment from the classification data with an object detection model and for generating garment segment data (Chen A, ¶0188, “the garment geometry and physics reconstruction module 118 may be configured to apply a 2.5 D multi-view garment physics simulation on the multi-projections of the 3D body avatar model [e.g. T. Adeyoola, N. Brown, N. Trott, E. Herbert, D. Robertson, J. Downing, N. Day, R. Boland, T. Boucher, J. Townsend, E. Clay, T. Warren, and A. Unadkat: “Computer implemented methods and systems for generating virtual body models for garment fit visualisation”. UK Patent GB2488237, Metail Limited, August 2012.].” ¶0189, “the garment metadata 134 may include garment attributes of style, colour, texture, shape, fit styles, vertical drapes (e.g. cuff, hem height, waist line position) material stretch and strain, etc” ¶0191, “The labelled garment image dataset 122 is a structured dataset containing a volume of garment image data and corresponding manually-generated semantic label map (indicating the regions of garment, skin, hair, background, etc.), attributes (including garment category, pattern, colour, material properties, camera view, etc.), control point/spline annotations (including garment landmark, body joints, skeleton layout, etc). The labelled garment image dataset 122 may be used as training data for learning machine learning models for garment segmentation, automatic landmark/control spline generation, body pose detection, camera view classification, etc.”),
a landmark extraction unit configured to receive the garment segment data and for detecting and extracting at least landmark feature present within the garment segment data and for generating landmark data (Chen A, ¶0098, “the data processing arrangement automatically predicts a visibility of and locations of one or more garment landmarks and/or splines on the model within the image.” ¶0099-0101, “the landmark and control spline prediction module may be configured to localise the landmarks characterising key positions on a garment. For example, the landmarks may correspond to the corners of the garment, positions of the armpit, etc. Further, the landmark and control spline prediction module may be configured to predict control splines (e.g. in the form of Bezier splines) which may be used to indicate the layer orders of garments in an outfit.”), and
a pattern generation unit configured to receive the landmark data and for generating in response two-dimensional garment pattern data (Chen A, ¶0169, “the landmark and control spline prediction module 108 may be configured to predict control splines (e.g. in the form of Bezier splines) which may be used to indicate the layer orders of garments in an outfit. This module is important for initializing finite-element garment physics simulation.”).
As to claim 5, claim 4 is incorporated and the combination of Chen A, Wang and Chen B discloses the material determination unit comprises a texture map determination unit for applying to the garment related data a semantic segmentation technique to generate one or more texture maps associated with the garment (Chen A, ¶0090, “The garment cut-out and layer dividing module may perform semantic segmentation of garment product photo using a deep neural network implemented using DeepLab” ¶0189, “The garment appearance data 132 may include, but is not limited, to cut-out garment texture sprites, texture maps for 3D mesh generated by back-projecting and stitching the 2D texture from each view, and alpha maps capturing the translucency of the garment.“ ¶0191, “The labelled garment image dataset 122 may be used as training data for learning machine learning models for garment segmentation”, and
a parameter determination unit for extracting from the garment textual data one or more fabric parameters associated with a fabric of the garment (Chen A, ¶0189, “the garment metadata 134 may include garment attributes of style, colour, texture, shape, fit styles, vertical drapes (e.g. cuff, hem height, waist line position) material stretch and strain, etc.” ¶0191, “attributes (including garment category, pattern, colour, material properties, camera view, etc.)” Wang, page 3, “our system automatically infers the parameters of the garment and the body from an input sketch and maps them to the nal 3D draped shape of the garment.” Fig. 5, “Each garment is parameterized by a small number of 2D sewing pattern dimensions as well as material properties including stretch, blend, and shear stiffness parameters.”),
wherein one or more of the texture maps and the fabric parameters form part of the virtual garment data (Chen A, ¶0189, “The garment appearance data 132 may include, but is not limited, to cut-out garment texture sprites, texture maps for 3D mesh generated by back-projecting and stitching the 2D texture from each view, and alpha maps capturing the translucency of the garment. The garment metadata 134 may include landmark points and control splines, which are useful for instructing physics simulation and layer composition in an outfit. Further, the garment metadata 134 may include garment attributes of style, colour, texture, shape, fit styles, vertical drapes (e.g. cuff, hem height, waist line position) material stretch and strain, etc.”).
As to claim 6, claim 5 is incorporated and the combination of Chen A, Wang and Chen B discloses the simulation engine comprises a tailor unit for receiving the virtual garment data and encoding the virtual garment data with the garment image data to form encoded garment data (Chen A, ¶0105, “The first stage of the two-stage algorithm includes using performing deep neural network coarse segmentation on a downsized image (such as 512×512 size image). As a method of implementation, the deep neural network models (e.g. SegNet [e.g. V. Badrinarayanan, A. Kendall, and R. Cipolla: “Segnet: A deep convolutional encoder-decoder architecture for image segmentation”. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017], DeepLab [e.g. L. Chen, G. Papandreou, L. Kokkinos, K. Murphy, and A. L. Yuille: “DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs”, arXiv preprint, 2016] may be trained using the product photos in Red Green Blue (RGB) or Red Green Blue Alpha (RGBA) formats as input and corresponding manually-labelled semantic label image as outputs.”),
a solver unit for receiving the encoded garment data and generating based thereon the simulated garment data (Wang, Page 2, “We achieve this by jointly training multiple encoder-decoder networks that each specializes at linking pairs of representations (e.g., recovering garment and body parameters from a sketch or recovering the on-body shapes of garments from the parameters) while operating at a common embedding. To train these network, we create a large scale synthetic dataset. Specically, we first define a set of parameterized garment types (shirt, skirt, and kimono) and generate different garments by sampling this representation. Then, we simulate each garment on a set of 3D body shapes and poses sampled from a deformable body model. Finally, for each of these simulated examples, we generate 2D sketches using non-photorealistic rendering. Thus, each examplar triplet in our dataset includes (i) a 2D sketch, (ii) garment and body parameters, and (iii) the resultant draped 3D shape of the garment. Subsequently, by jointly training multiple encoder-decoder networks via a novel multimodal loss function, we learn a common embedding that can be queried using any of the different modalities.”), and
a feedback loop for feeding back the simulated garment data to the tailor unit for further processing thereby (Wang, Page 5, “(i) sketch descriptor to latent space (FS2L = S → L), (ii) parameter space to latent space (FP2L = P → L), (iii) latent space to parameter space (FL2P = L → P), and (iv) latent space to the draped garment shape (FL2M = L → M). We learn these mappings by jointly training four encoder-decoder neural networks (i.e., sketch-to-parameter, sketch-to-3D garment shape, parameter to-3D garment shape, and parameter-to-parameter) that share a common embedding space (see Figure 6). We describe the network architecture with more details in section 4.”).
As to claim 7, claim 5 is incorporated and the combination of Chen A, Wang and Chen B discloses the simulation engine comprises a one shot network for receiving the virtual garment data and the avatar data and for generating based thereon the simulated garment data (Chen A, ¶0087, “simulating digital garment models on customer's body shape/pose, compositing different digital garment models into an outfit, and rendering a dressed avatar of the customer for final visualization. Therefore, the second subsystem simulates the interactive garment model onto a 3D body avatar of a consumer.” Chen B, abstract, “compresses DNNs into slimmer architectures with competitive performances and significant FLOPs reductions by Only-Train-Once (OTO).”).
As to claim 8, claim 7 is incorporated and the combination of Chen A, Wang and Chen B discloses the one-shot network comprises a convolutional neural network (Chen B, Page 7, “we numerically demonstrate the effectiveness of OTO by one-shot training and pruning without fine-tuning on several benchmark compression tasks for CNNs”).
As to claim 9, claim 1 is incorporated and the combination of Chen A, Wang and Chen B discloses a feedback loop for feeding back the garment related training data to the tailor unit for further processing thereby (Wang, Page 5, “(i) sketch descriptor to latent space (FS2L = S → L), (ii) parameter space to latent space (FP2L = P → L), (iii) latent space to parameter space (FL2P = L → P), and (iv) latent space to the draped garment shape (FL2M = L → M). We learn these mappings by jointly training four encoder-decoder neural networks (i.e., sketch-to-parameter, sketch-to-3D garment shape, parameter to-3D garment shape, and parameter-to-parameter) that share a common embedding space (see Figure 6). We describe the network architecture with more details in section 4.”).
As to claim 10, claim 1 is incorporated and the combination of Chen A, Wang and Chen B discloses the avatar generation unit comprises a segmentation unit for processing the user image data and generating silhouette image data (Chen A, ¶0118-0119, “the data processing arrangement receives inputs from the customer, wherein the inputs include one or more consumer body metrics, and the second subsystem is configured to generate automatically a 3D body avatar representing the consumer's body shape.” ¶0181, “The body pose matching and pose changing module 114 may be configured to detect the underlying body pose in an input image and adjust the pose of a 3D virtual body model to match that in the input image.”), and
a machine learning unit for receiving the silhouette image data and the user biometric data to form input data, and for applying to the input data one or more machine learning models to generate the avatar (Chen A, ¶0119, “the second subsystem creates the 3D body avatar representing the consumer's body shape using a 3D statistical body model with an embedded skeleton and a regression application. In some embodiments, the body shape estimation module may generate a full 3D geometry of a body avatar from body measurements and parameters provided by the customer in the UI (e.g. height, weight, cup size, bust circumference, waist circumference, hips circumference, inside leg length, etc.).” ¶0121, “the second subsystem simulates automatically a drape and a fit of the digital garment model onto the 3D body avatar. The second subsystem may simulate the drape and the fit of the digital garment model onto the 3D body avatar using finite-element analysis. In some embodiments, the garment physics simulation module may be configured to perform finite-element analysis of the digital garment model(s) on the generated 3D body avatar of the customers. At the runtime, the garment physics simulation module models how the garment model will drape and fit against, or on, a customer body shape.”).
As to claim 11, claim 10 is incorporated and the combination of Chen A, Wang and Chen B discloses the machine learning model comprises a deep learning regression model, and wherein the machine learning unit further generates a plurality of measurements forming part of the avatar related data by employing a learned mapping technique by mapping between a predefined human silhouette and a plurality of body measurements of the user, and applies a statistical anthropomorphic analysis technique to the user biometric data (Chen A, ¶0119, “the data processing arrangement renders the consumer's body metrics selectable from height, weight, cup size, bust circumference, waist circumference, hips circumference, inside leg length. Furthermore, the second subsystem creates the 3D body avatar representing the consumer's body shape using a 3D statistical body model with an embedded skeleton and a regression application. In some embodiments, the body shape estimation module may generate a full 3D geometry of a body avatar from body measurements and parameters provided by the customer in the UI (e.g. height, weight, cup size, bust circumference, waist circumference, hips circumference, inside leg length, etc.). The body shape estimation module may be implemented based on a 3D statistical body model with skeleton embedded and applying a regression to estimate intrinsic model parameters from input measurements for 3D reconstruction [e.g. T. Adeyoola, N. Brown, N. Trott, E. Herbert, D. Robertson, J. Downing, N. Day, R. Boland, T. Boucher, J. Townsend, E. Clay, T. Warren, and A. Unadkat: “Computer implemented methods and systems for generating virtual body models for garment fit visualisation”. UK Patent GB2488237, Metail Limited, August 2012].”.
As to claim 12, claim 1 is incorporated and the combination of Chen A, Wang and Chen B discloses a fit analysis unit for receiving the virtual garment data and the avatar data and for generating in response thereto garment fit data indicative of a fit of the virtual garment on the avatar (Chen A, ¶0121, “At the runtime, the garment physics simulation module models how the garment model will drape and fit against, or on, a customer body shape.” ¶0188, “the garment geometry and physics reconstruction module 118 may be configured to apply a 2.5 D multi-view garment physics simulation on the multi-projections of the 3D body avatar model [e.g. T. Adeyoola, N. Brown, N. Trott, E. Herbert, D. Robertson, J. Downing, N. Day, R. Boland, T. Boucher, J. Townsend, E. Clay, T. Warren, and A. Unadkat: “Computer implemented methods and systems for generating virtual body models for garment fit visualisation”. UK Patent GB2488237, Metail Limited, August 2012.].” ¶0204, “the garment model 1712 may be fitted on the 3D model 1716 via a visualization engine 1718 to provide a final output 1720. The visualization engine 1718 may include the garment physics simulation module 1606 and the outfitting and rendering module 1608.”).
As to claim 13, claim 12 is incorporated and the combination of Chen A, Wang and Chen B discloses the virtual garment data is indicative of a virtual garment, and wherein the fit analysis unit includes a processor programmed to:
segment the virtual garment into a plurality of garment segments (Chen A, ¶0042, “segment an image of the garment to produce a set of segmented garment images” ¶0090, “The garment cut-out and layer dividing module may perform semantic segmentation of garment product photo using a deep neural network implemented using DeepLab [e.g. L. Chen, G. Papandreou, L. Kokkinos, K. Murphy, and A. L. Yuille, “DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs”, arXiv preprint, 2016” ¶0105, “the garment cut-out and layer dividing module may be used to split further the garment texture into one or more layers corresponding to one or more body parts. For example, a leg wear may be split into two layers comprising a left-leg layer and a right-leg layer. The garment cut-out and layer dividing module may use a two-stage algorithm to split the garment texture into one or more layers.” ¶0163, ‘The output photo 604 with segmented garment textures 606-608 may be obtained after the input product model photo 602 is processed by the garment cut-out and layer dividing module 106.”), and
classify each of the plurality of segments into one or more garment related categories (Chen A, ¶0105, “the garment cut-out and layer dividing module may be used to split further the garment texture into one or more layers corresponding to one or more body parts. For example, a leg wear may be split into two layers comprising a left-leg layer and a right-leg layer. The garment cut-out and layer dividing module may use a two-stage algorithm to split the garment texture into one or more layers.” ¶0163, “The output photo 604 with segmented garment textures 606-608 may be obtained after the input product model photo 602 is processed by the garment cut-out and layer dividing module 106. Furthermore, accessories such as shoes 610 are also segmented. Yet further, the body of the model (arms and hands 612 and legs 614) is also segmented. Moreover, the background 616 in the input product model photo 602 is also segmented. In an embodiment, the input product model photo 602 may be processed by DeepLab [e.g. L. Chen, G. Papandreou, L. Kokkinos, K. Murphy, and A. L. Yuille: “DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs”, arXiv preprint, 2016] to obtain the segmented output photo 604.”), and
generate the garment fit data based on the one or more garment related categories (Chen A, ¶0121, “simulates automatically a drape and a fit of the digital garment model onto the 3D body avatar. The second subsystem may simulate the drape and the fit of the digital garment model onto the 3D body avatar using finite-element analysis. In some embodiments, the garment physics simulation module may be configured to perform finite-element analysis of the digital garment model(s) on the generated 3D body avatar of the customers. At the runtime, the garment physics simulation module models how the garment model will drape and fit against, or on, a customer body shape.” ¶0133, “texture attributes, shape attributes, fit styles attributes, vertical drapes attributes, cuff, hem height, waist line position”).
As to claim 14, claim 13 is incorporated and the combination of Chen A, Wang and Chen B discloses the processor is further programmed to generate a texture map indicative of the fit of the garment on the avatar (Chen A, ¶0079, “garment texture alignment and sprite scale estimation” ¶0091, “the garment texture in-painting module may be used to semi-automatically complete any occluded garment texture (or occluded sections) in a product photo using an interactive UI. The occluded textures may result due to occluding hair or arms of the model. Accordingly, the garment texture in-painting module may fill up the occluded texture either by cloning the texture samples or by interpolating the colours from neighbouring unoccluded areas. A goal of such a method is to generate an overall visually plausible and coherent texture of the whole garment.” ¶0105, “split further the garment texture into one or more layers corresponding to one or more body parts.” ¶0114, “apply a spatial transformation on the cut-out garment texture sprites so that it may be laid on the proper positions on the standard body avatar used for garment digitisation.” ¶0143, “cut-out garment texture sprites, texture maps for a 3D mesh”).
As to claim 15, the combination of Chen A, Wang and Chen B discloses a method for generating a rendered virtual image of a garment on an avatar of a user, comprising generating, in response to receiving garment related data, virtual garment data with a virtual garment generation unit, wherein the garment related data includes garment image data and garment textual data,generating with an avatar generation unit, in response to receiving user data, user avatar data that is representative of the avatar of the user, wherein the user data includes at least one of user image data and user biometric data,generating with a simulation engine, in response to receiving the user avatar data and the virtual garment data, simulated garment data representative of a simulated garment draped on the avatar, andrendering the simulated garment data with a rendering unit and generating the rendered virtual image,wherein generating with the simulation engine includes encoding with a tailor unit the virtual garment data with the garment image data to form encoded garment data,providing a graph network configured for receiving the encoded garment data and for generating garment related training data, andproviding a one-shot network configured for receiving the garment related training data in order to train the one-shot network, and wherein the one-shot network is configured for receiving the virtual garment data and the user avatar data and for generating based thereon the simulated garment data (See claim 1 for detailed analysis.).
As to claim 16, claim 15 is incorporated and the combination of Chen A, Wang and Chen B discloses generating the virtual garment data with the virtual garment generation unit comprises generating with a pattern determination unit, in response to receiving the garment image data and the garment textual data, two-dimensional garment pattern data based on the garment image and the garment textual data,generating with a mesher unit, in response to receiving and processing the two dimensional garment pattern data, a plurality of three-dimensional garment meshes forming at least part of the virtual garment data, anddetermining with a material determination unit a type of material associated with the garment based on the garment image data and the garment textual data (See claim 2 for detailed analysis.).
As to claim 17, claim 16 is incorporated and the combination of Chen A, Wang and Chen B discloses the mesher unit employs one or more meshing techniques, comprising subdividing with the meshing technique the two-dimensional garment pattern data into the plurality of three-dimensional garment meshes (See claim 3 for detailed analysis.).
As to claim 18, claim 17 is incorporated and the combination of Chen A, Wang and Chen B discloses the garment related data is indicative at least one garment, and wherein generating with the pattern determination unit comprises classifying the garment into one of a plurality of different categories indicative of different types of garments and for generating in response garment classification data,detecting and extracting at least one garment segment from the garment classification data with an object detection model and for generating garment segment data,detecting and extracting at least landmark feature present within the garment segment data and for generating landmark data, andgenerating in response to the landmark data two-dimensional garment pattern data (See claim 4 for detailed analysis.).
As to claim 19, claim 18 is incorporated and the combination of Chen A, Wang and Chen B discloses determining with the material determination unit comprises applying to the garment related data a semantic segmentation technique to generate one or more texture maps associated with the garment, andextracting from the garment textual data one or more fabric parameters associated with a fabric of the garment,wherein one or more of the texture maps and the fabric parameters form part of the virtual garment data (See claim 5 for detailed analysis.).
As to claim 20, claim 19 is incorporated and the combination of Chen A, Wang and Chen B discloses generating with the simulation engine comprises providing a tailor unit for receiving the virtual garment data and encoding the virtual garment data with the garment image data to form encoded garment data,providing a solver unit for receiving the encoded garment data and generating based thereon the simulated garment data, andfeeding back the simulated garment data to the tailor unit for further processing thereby (See claim 6 for detailed analysis.).
As to claim 21, claim 19 is incorporated and the combination of Chen A, Wang and Chen B discloses generating with the simulation engine comprises employing a one shot network for receiving the virtual garment data and the avatar data and for generating based thereon the simulated garment data (See claim 7 for detailed analysis.).
As to claim 22, claim 15 is incorporated and the combination of Chen A, Wang and Chen B discloses feeding back with a feedback loop the garment related training data to the tailor unit for further processing thereby (See claim 9 for detailed analysis.).
As to claim 23, claim 15 is incorporated and the combination of Chen A, Wang and Chen B discloses generating with the avatar generation unit comprises generating, in response to receiving the user image data, silhouette image data, andapplying, in response to receiving the silhouette image data and the user biometric data, one or more machine learning models to generate the avatar (See claim 10 for detailed analysis.).
As to claim 24, claim 23 is incorporated and the combination of Chen A, Wang and Chen B discloses the machine learning model comprises a deep learning regression model, comprising generating a plurality of measurements forming part of the avatar related data by employing a learned mapping technique by mapping between a predefined human silhouette and a plurality of body measurements of the user, and applies a statistical anthropomorphic analysis technique to the user biometric data (See claim 11 for detailed analysis.).
As to claim 25, claim 15 is incorporated and the combination of Chen A, Wang and Chen B discloses generating with a fit analysis unit, in response to receiving the virtual garment data and the avatar data, garment fit data indicative of a fit of the virtual garment draped on the avatar (See claim 12 for detailed analysis.).
As to claim 26, claim 25 is incorporated and the combination of Chen A, Wang and Chen B discloses the virtual garment data is indicative of a virtual garment, further comprising segmenting the virtual garment into a plurality of garment segments, classifying each of the plurality of segments into one or more garment related categories, and generating the garment fit data based on the one or more garment related categories (See claim 13 for detailed analysis.).
As to claim 27, claim 26 is incorporated and the combination of Chen A, Wang and Chen B discloses generating a texture map indicative of the fit of the garment on the avatar (See claim 14 for detailed analysis.).
As to claim 28, the combination of Chen A, Wang and Chen B discloses a non-transitory, computer readable medium comprising computer program instructions tangibly stored on the computer readable medium, wherein the computer program instructions are executable by at least one computer processor to perform a method, the method comprising:generating, in response to receiving garment related data, virtual garment data with a virtual garment generation unit, wherein the garment related data includes garment image data and garment textual data,generating with an avatar generation unit, in response to receiving user data, user avatar data that is representative of the avatar of the user, wherein the user data includes at least one of user image data and user biometric data,generating with a simulation engine, in response to receiving the user avatar data and the virtual garment data, simulated garment data representative of a simulated garment draped on the avatar, andrendering the simulated garment data with a rendering unit and generating the rendered virtual image wherein generating with the simulation engine includes encoding with a tailor unit the virtual garment data with the garment image data to form encoded garment data,providing a graph network configured for receiving the encoded garment data and for generating garment related training data, andproviding a one-shot network configured for receiving the garment related training data in order to train the one-shot network, and wherein the one-shot network is configured for receiving the virtual garment data and the user avatar data and for generating based thereon the simulated garment data (See claim 1 for detailed analysis.).
As to claim 29, claim 28 is incorporated and the combination of Chen A, Wang and Chen B discloses generating the virtual garment data with the virtual garment generation unit, comprises generating with a pattern determination unit, in response to receiving the garment image data and the garment textual data, two-dimensional garment pattern data based on the garment image and the garment textual data,generating with a mesher unit, in response to receiving and processing the two dimensional garment pattern data, a plurality of three-dimensional garment meshes forming at least part of the virtual garment data, anddetermining with a material determination unit a type of material associated with the garment based on the garment image data and the garment textual data (See claim 2 for detailed analysis.).
As to claim 30, claim 29 is incorporated and the combination of Chen A, Wang and Chen B discloses subdividing with a meshing technique the two-dimensional garment pattern data into the plurality of three- dimensional garment meshes (See claim 3 for detailed analysis.).
As to claim 31, claim 28 is incorporated and the combination of Chen A, Wang and Chen B discloses the garment related data is indicative at least one garment, and wherein generating with the pattern determination unit comprisesclassifying the garment into one of a plurality of different categories indicative of different types of garments and for generating in response garment classification data,detecting and extracting at least one garment segment from the garment classification data with an object detection model and for generating garment segment data,detecting and extracting at least landmark feature present within the garment segment data and for generating landmark data, andgenerating in response to the landmark data two-dimensional garment pattern data (See claim 4 for detailed analysis.).
As to claim 32, claim 31 is incorporated and the combination of Chen A, Wang and Chen B discloses determining with the material determination unit, comprises applying to the garment related data a semantic segmentation technique to generate one or more texture maps associated with the garment, andextracting from the garment textual data one or more fabric parameters associated with a fabric of the garment,wherein one or more of the texture maps and the fabric parameters form part of the virtual garment data (See claim 5 for detailed analysis.).
As to claim 33, claim 28 is incorporated and the combination of Chen A, Wang and Chen B discloses generating with the avatar generation unit comprises generating, in response to receiving the user image data, silhouette image data, andapplying, in response to receiving the silhouette image data and the user biometric data, one or more machine learning models to generate the avatar (See claim 10 for detailed analysis.).
As to claim 34, claim 33 is incorporated and the combination of Chen A, Wang and Chen B discloses generating a plurality of measurements forming part of the avatar related data by employing a learned mapping technique by mapping between a predefined human silhouette and a plurality of body measurements of the user, and applies a statistical anthropomorphic analysis technique to the user biometric data (See claim 11 for detailed analysis.).
As to claim 35, claim 28 is incorporated and the combination of Chen A, Wang and Chen B discloses comprising generating with a fit analysis unit, in response to receiving the virtual garment data and the avatar data, garment fit data indicative of a fit of the virtual garment draped on the avatar (See claim 12 for detailed analysis.).
As to claim 36, claim 35 is incorporated and the combination of Chen A, Wang and Chen B discloses the virtual garment data is indicative of a virtual garment, further comprising segmenting the virtual garment into a plurality of garment segments, classifying each of the plurality of segments into one or more garment related categories, and generating the garment fit data based on the one or more garment related categories (See claim 13 for detailed analysis.).
As to claim 37, claim 36 is incorporated and the combination of Chen A, Wang and Chen B discloses generating a texture map indicative of the fit of the garment on the avatar (See claim 14 for detailed analysis.).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Chen et al. (US Pub 2020/0320769 A1) discloses mapping the stored extracted attributes of the garment to physical fabric properties of the garment, and/or to model parameters for garment physics simulation.
Gundogdu, E., Constantin, V., Seifoddini, A., Dang, M., Salzmann, M., & Fua, P. (2019, October). (Garnet: A two-stream network for fast and accurate 3d cloth draping. In 2019 IEEE/CVF international conference on computer vision (ICCV) (pp. 8738-8747). IEEE.) teaches a two-stream network architecture that can drape a 3D garment shape on
different target bodies in many different poses, while running 100 times faster than a physics-based simulator.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to YU CHEN whose telephone number is (571)270-7951. The examiner can normally be reached on M-F 8-5 PST Mid-day flex.
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, Xiao Wu can be reached on 571-272-7761. 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.
/YU CHEN/Primary Examiner, Art Unit 2613
Prosecution Insights