Generating augmented reality pre-renderings using a template image

Abstract

The system and method for generating augmented reality pre-renderings can provide the benefits of augmented reality rendering without requiring the use of user data. Template images can be used in place of user data to protect user privacy and allow users to view objects or products rendered on one or a variety of template images of their choice.

Description

[Technical field]

[0001] Related Applications This application claims priority to and the benefit of U.S. Non-provisional Patent Application No. 17 / 153,263, filed January 20, 2021. U.S. Non-provisional Patent Application No. 17 / 153,263 is incorporated herein by reference in its entirety.

[0002] The present disclosure relates generally to pre-rendering images and, more particularly, to pre-rendering augmented reality images from a set of template images to enable a limited augmented reality experience from the template images, for example, as an alternative to a real-time personalized augmented reality experience. [Background technology]

[0003] Augmented reality (AR) can refer to the creation and execution of interactive experiences of real-world environments in which objects present in the real world are augmented with computer-generated perceptual information. As an example, an AR experience can include augmenting a scene captured by a camera by inserting virtual objects into the scene and / or modifying the appearance of real-world objects included in the scene.

[0004] When searching for aesthetically-intense products, such as cosmetics, often it is not enough to simply look at the product packaging, or even the product itself. To solve this problem, efforts have been made to digitize cosmetics and other products in augmented reality (AR) to allow consumers to visualize the product on themselves or within their personal environment. However, many users may not want to use AR try-on due to natural friction. For example, the user may not be in a location where using a camera is practical, the user may not feel like they look their best, may be reluctant to turn on the camera, and / or the user may simply not want to grant camera permission.

[0005] Furthermore, live AR experiences can require large data bandwidth and processing power. Summary of the Invention [Means for solving the problem]

[0006] Aspects and advantages of embodiments of the disclosure will be set forth in part in the description that follows, or may be learned from the description, or may be learned by practice of the embodiments.

[0007] One exemplary aspect of the present disclosure is directed to a computer-implemented method for providing pre-rendered augmented images. The method can include obtaining, by a computing device, a plurality of template images. The method can include processing, by the computing device, the plurality of template images with an augmented reality rendering model to generate a plurality of pre-rendered images. In some implementations, the method can include receiving, by the computing device, a request and preferences for a result image. The method can include providing, by the computing device, a pre-rendered result based at least in part on the request and preferences. In some implementations, the pre-rendered result can be a pre-rendered image from a plurality of pre-rendered images.

[0008] Another exemplary aspect of the present disclosure is directed to a computing system. The computing system may include one or more processors and one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors, cause the computing system to perform operations. The operations may include obtaining an augmented reality asset. In some implementations, the augmented reality asset may include digitization parameters. The operations may include obtaining a plurality of template images. The operations may include processing the plurality of template images with the augmented reality model to generate a plurality of pre-rendered images based at least in part on the digitization parameters, and storing the plurality of pre-rendered images.

[0009] Another exemplary aspect of the present disclosure is directed to one or more non-transitory computer-readable media that collectively store instructions that, when executed by one or more processors, cause a computing system to perform operations. The operations can include obtaining an augmented reality asset. The augmented reality asset can include digitization parameters. In some implementations, the operations can include obtaining a plurality of template images. The operations can include processing the plurality of template images with the augmented reality model to generate a plurality of pre-rendered images based at least in part on the digitization parameters. The operations can include storing the plurality of pre-rendered images on a server and receiving a search query that can include one or more search terms. The one or more search terms can be related to a product. In some implementations, the operations can include providing search results. The search results can include pre-rendered images from the plurality of pre-rendered images retrieved from the server. The pre-rendered images can include renderings of the product.

[0010] Other aspects of the present disclosure are directed to various systems, apparatus, non-transitory computer-readable media, user interfaces, and electronic devices.

[0011] These and other features, aspects, and advantages of various embodiments of the present disclosure will become better understood with reference to the following description and appended claims. The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate exemplary embodiments of the present disclosure and, together with the description, serve to explain the relevant principles.

[0012] Detailed descriptions of embodiments, directed to those skilled in the art, are set forth herein with reference to the accompanying drawings. [Brief description of the drawings]

[0013] [Figure 1A] FIG. 1 is a block diagram of an exemplary computing system that performs pre-rendering, according to an exemplary embodiment of the present disclosure. [Figure 1B] FIG. 2 is a block diagram of an exemplary computing device that performs pre-rendering, according to an exemplary embodiment of the present disclosure. [Figure 1C] FIG. 2 is a block diagram of an exemplary computing device that performs pre-rendering, according to an exemplary embodiment of the present disclosure. [Diagram 2] FIG. 2 is a block diagram of an exemplary display of a pre-rendered image, according to an exemplary embodiment of the present disclosure. [Diagram 3] FIG. 2 is a block diagram of an exemplary display of a pre-rendered image, according to an exemplary embodiment of the present disclosure. [Figure 4] FIG. 2 is a block diagram of an exemplary display of a pre-rendered image, according to an exemplary embodiment of the present disclosure. [Diagram 5] FIG. 1 is a block diagram of an exemplary pre-rendering system, in accordance with an exemplary embodiment of the present disclosure. [Figure 6] FIG. 2 is a flowchart diagram of an example method for performing pre-rendering, according to an example embodiment of the present disclosure. [Figure 7] FIG. 2 is a flowchart diagram of an example method for performing pre-rendering, according to an example embodiment of the present disclosure. [Figure 8] FIG. 2 is a flowchart diagram of an example method for performing pre-rendering, according to an example embodiment of the present disclosure. [Figure 9] FIG. 1 is a block diagram of an exemplary pre-rendering system, in accordance with an exemplary embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] Reference numbers that are repeated among the drawings are intended to identify like features in various implementations.

[0015] Overview In general, the present disclosure is directed to systems and methods for pre-rendering an augmented reality template image to enable a limited augmented reality experience from the template image, for example, as an alternative to a real-time personalized augmented reality experience using a user's specific image. In some implementations, the system and method can be implemented as a platform for generating pre-rendered augmented reality images from a set of template images. The system can include obtaining augmented reality assets and a set of template images. The template images can be, for example, a diverse set of images showing people or settings with diverse characteristics or features, such as different eye colors or other visual characteristics. The set of template images can be processed by an augmented or augmented reality model parameterized by the obtained augmented reality assets. The augmented reality model can output a set of pre-rendered images, and the pre-rendered images can be stored on a server. In some implementations, the system can receive a request for result images along with user preferences. For example, the request can be a search query in the form of one or more search terms. The system can process the request and preferences to determine a result image. The result image can be one or more pre-rendered images from the set of pre-rendered images. For example, a user can select (or have preselected) one or more preferences, which can guide the selection of one or more pre-rendered images as result images. The results can be provided to the user. In that way, the user can be provided with one or more pre-rendered images that satisfy the user's preferences (e.g., pre-rendered images generated from templates having characteristics that match the user's characteristics). Thus, in some cases, a user can see how a product would appear on a person or setting similar to themselves, but without having to be in a location where the use of a camera is practical and / or without having to supply actual images of themselves or their environment.In some implementations, the system can also provide the user with a link to the real-time augmented reality experience, for example, so that the user can continue to pursue the full real-time AR experience if desired.

[0016] In some implementations, the pre-rendering of the augmented reality images can be facilitated by a platform. The platform can be used to collect augmented reality data assets created using the platform, or alternatively, collect augmented reality assets generated outside the platform. The platform can collect augmented reality assets from third-party companies offering products for sale (e.g., cosmetics (e.g., lipstick, eye shadow, etc.), furniture or other household items (e.g., electronics, cookware, glassware, ornaments, plants, etc.), clothing, paint colors, automobiles, various electronic devices, etc.). Additionally, the platform can use the collected augmented reality assets to render rendering effects in template images to generate pre-rendered images. The template images can be stored locally or obtained from outside the platform. The pre-rendered images can then be stored for later display to the user. For example, a user computing device can send a request for one or more pre-rendered images. The request can be a search query, a user selection, or an automated request in response to a user action. The platform can process the request along with preferences to provide one or more pre-rendered images related to the request and the user preferences.

[0017] The systems and methods can include obtaining a plurality of template images. The plurality of template images can be a set of images depicting a similar focus but with differences (e.g., each of the template images depict a living room, but each photo depicts a different color scheme for the room, with furniture and décor having different themes and colors). The systems and methods can process the plurality of template images with an augmented reality rendering model to generate a plurality of pre-rendered images. In some implementations, the augmented reality rendering model can include object tracking and rendering. The plurality of pre-rendered images can be a set of augmented images, where each of the template images can be augmented to include augmented reality rendering effects. The augmented reality rendering effects can be products sold by a third party, where a consumer can virtually "try on" the product in different template scenarios (e.g., furniture items rendered in the template room images). In some implementations, the systems and methods can include receiving a request and a preference for a result image. The request can come from a user computing system, where the user indicates in the request a desire to see pre-rendered images. In some implementations, the request can include search terms entered into a search engine. The preferences can include selections made at the time of the request or previously stored that can indicate a preferred template type or a preferred template. The systems and methods can then provide a pre-rendered result based at least in part on the request and the preferences, where the pre-rendered result can be a pre-rendered image from a plurality of pre-rendered images. In some implementations, the pre-rendered result can be a pre-rendered image that matches the preferences, where the preferences include a template selected by the user from a plurality of template images.In some implementations, the systems and methods may provide a link to an augmented reality rendering experience for live try-on.

[0018] In some implementations, the system and method can obtain an augmented reality asset, which can be stored for use by the augmented reality model or augmented model. The augmented reality asset can include digitization parameters. The digitization parameters can enable the augmented reality model or augmented model to render a particular rendering effect. In some implementations, the augmented reality asset can be used by the augmented reality model to process a set of template images to generate a set of pre-rendered images. The set of pre-rendered images can then be stored for later retrieval. In some implementations, the set of pre-rendered images can be stored on a server.

[0019] In some implementations, the set of template images can be processed by object tracking computer vision algorithms and computer rendering operations to generate pre-rendered augmented reality images that mimic the appearance of the product. The product can be applied to a depicted face in the template images or inserted into a depicted image of the set of template images. The rendering operations can be influenced by CPU or GPU algorithms and corresponding parameters. The corresponding parameters can be parameters that realistically capture the appearance of the product inserted into the template images under depicted lighting conditions.

[0020] In some implementations, the systems and methods can include receiving a search query including one or more search terms, the one or more search terms relating to the product. Additionally, the systems and methods can include providing search results. The search results can include a pre-rendered image from a plurality of pre-rendered images. Additionally, the pre-rendered image can include a rendering of the product.

[0021] In some implementations, the set of template images can be processed to generate a set of template models that can be processed by the augmented reality model or the augmented model. In some implementations, the template models can be modified before being processed. The systems and methods can receive input to modify a template model of the plurality of template models. The systems and methods can modify the template model based at least in part on the input. In some implementations, the systems and methods can include providing a template model of the plurality of template models for display.

[0022] In some implementations, the augmented reality model may include a perception subgraph and a rendering subgraph. The perception subgraph may be uniform across the system. The perception subgraph may be used with a variety of different rendering subgraphs. The rendering subgraph may be built by a third party to generate the rendering effect provided to the user. The rendering subgraph may be built and then used by an augmented reality pre-rendering platform that stores the perception subgraph. The rendering subgraph may vary depending on the rendering effect and the third party. In some implementations, a single perception subgraph may be used with multiple rendering subgraphs to render multiple renderings in an augmented image or video. For example, a photo or video of a face may be processed to generate an augmented reality rendering of lipstick, eyeshadow, and mascara on the face. The processing may include a single perception subgraph, but may include rendering subgraphs for each product (i.e., lipstick, eyeshadow, and mascara).

[0023] The systems and methods disclosed herein can be applicable to a variety of augmented reality experiences (e.g., household goods, makeup, automotive (3D), glasses, jewelry, clothing, and haircuts). For example, the systems and methods disclosed herein can be used to generate a set of pre-rendered images to provide consumers with renderings of products in various environments or applications. In some implementations, the template images can be various room images with different setups, color schemes, and decorations. The desired rendering effect can be a sofa. The system can populate the set of template images and augmented reality assets related to the sofa and generate a set of pre-rendered images to provide to the user. The pre-rendered images can include an augmented reality rendering of a desired sofa in each of the template images to allow the user to see how the sofa might look in various room types with different color schemes, decorations, and layouts.

[0024] In some implementations, the multiple template images can include a set of face images having diverse characteristics or features, such as, for example, different eye colors or other visual characteristics, whereby the consumer can select and / or store a preferred template that may be most similar to the consumer. The rendering effect can be a makeup rendering (e.g., lipstick, eye shadow, mascara, foundation, etc.). The augmented reality model can process the augmented reality assets and the set of template images to generate a pre-rendered makeup image, where each of the multiple template images can be augmented to include a makeup product. For example, a particular shade of lipstick can be rendered within each of the template images.

[0025] In some implementations, a set of template images can be manually or automatically selected from a large corpus of images to provide a representative group of template images that show representative images of different scenarios or features for a given topic. For example, a topic can be a room, a neighbor, a face, etc.

[0026] The systems and methods disclosed herein can be used to provide personalized advertisements to users. For example, user preferences can be stored. The stored preferences allow the provided product advertisements to be adapted to render the advertised products on the user's preferred template images.

[0027] In some implementations, the augmented reality assets can be managed, generated, and / or reproduced by the product brand. The digitization parameters can be imported from third-party companies. In some implementations, the digitization parameters can be extracted from third-party rendering engines or exported from provided templates for augmented reality rendering effect generation.

[0028] Additionally, the perceptual model can be adjusted manually or automatically to provide an optimized mesh. The adjustments can be responsive to lighting or varying image quality. The platform can provide a preview to assist in modifying the augmented reality model. In some implementations, the platform can include a pipeline for retrieving images.

[0029] The pre-rendered images may include tagging to index the rendered product.

[0030] In some implementations, the indexed products may include a consistent naming scheme for products and product colors for better search results. The platform may include a data structure that can relate augmented reality assets to specific semantic or lexicographical entities. This data structure may be useful for understanding search queries to create product mappings.

[0031] In some implementations, the pre-rendered images may be provided as a carousel for the user to scroll through. Alternatively, various pre-rendered images with different product renderings but the same template image may be provided in the form of a carousel to allow for personalized previews. Additionally, in some implementations, the carousel may include a virtual "try-on" panel to provide the user with an augmented reality rendering experience that processes the user's data to provide user-augmented images or videos.

[0032] In some implementations, the augmented reality platform may retrieve data assets for rendering the augmented reality effects via a system and method for data asset acquisition. The system and method for data asset acquisition may involve one or more systems or devices. The first computing device may be a server, a facilitating computing device, or an intermediary computing device. The second computing device may be a third-party computing device. The third party may be a video game company, a product manufacturer, or a product brand. The first computing device and the second computing device may exchange data to generate an augmented reality rendering experience for the user. The augmented reality rendering experience may include rendering an augmented reality view including one or more products or items. The products may be cosmetics (e.g., lipstick, eye shadow, etc.), furniture or other household items (e.g., electronics, cookware, glassware, ornaments, plants, etc.), clothing, paint colors, automobiles, various electronic devices, or other items.

[0033] The acquisition of the data assets may include the first computing device sending a software development kit to the second computing device. The software development kit may include templates for constructing rendering effect shaders. The software development kit may include example effects, tools for constructing the rendering effect, and a preview mode useful for constructing the augmented reality rendering. The second computing device may be used to construct the rendering effect, and once the rendering effect is constructed, the second computing device may export the constructed rendering effect data to a renderable compressed file (e.g., a .ZIP file) that may include the data assets necessary to recreate the rendering effect. The data assets may then be sent to the first computing device. Upon receiving the data assets, the first computing device may store the data assets for use in an augmented reality rendering experience provided to a user. The provided augmented reality rendering experience may be provided to a user, where the user may input user data for processing, and the output may be augmented user data including the rendering effect constructed on the second computing device. The user data may be image data or video data captured by the user device. In some implementations, the user data may be a live camera feed.

[0034] Further, in some implementations, the system and method may be used as a visual compatibility calculator. For example, the system and method may be used to ensure that a particular product or part will fit a desired space or location. The system and method may be used to virtually test the dimensions / size of a product using virtual reality. A third party may provide a data asset that may include data describing the measurements of a product. The data asset may then be used to provide a user with an augmented reality rendering experience in which the product is rendered according to the measurements provided by the third party. This aspect allows a consumer to "try on" the product to visualize the space the product will occupy.

[0035] The systems and methods of the present disclosure provide many technical effects and advantages. As an example, the systems and methods can generate a set of pre-rendered images using a set of template images and an augmented reality experience. The systems and methods can also be further used to provide consumers with pre-rendered images of products on various templates. Additionally, the systems and methods can enable consumers to view pre-rendered augmented reality renderings on templates that meet the consumer's preferences.

[0036] Another technical advantage of the systems and methods of the present disclosure is their ability to provide "try-on" images that match consumer shopping preferences when an augmented reality real-time try-on experience is less than ideal.

[0037] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the drawings.

[0038] Exemplary Devices and Systems 1A illustrates a block diagram of an exemplary computing system 100 for performing pre-rendering, according to an exemplary embodiment of the present disclosure. The system 100 includes a user computing device 102, a server computing system 130, and a training computing system 150 communicatively coupled via a network 180.

[0039] The user computing device 102 may be any type of computing device, such as, for example, a personal computing device (e.g., a laptop or desktop), a mobile computing device (e.g., a smartphone or tablet), a game console or controller, a wearable computing device, an embedded computing device, or any other type of computing device.

[0040] The user computing device 102 includes one or more processors 112 and memory 114. The one or more processors 112 may be any suitable processing device (e.g., a processor core, a microprocessor, an ASIC, an FPGA, a controller, a microcontroller, etc.) and may be one processor or multiple processors operably connected. The memory 114 may include one or more non-transitory computer-readable storage media, such as RAM, ROM, EEPROM, EPROM, flash memory devices, magnetic disks, etc., and combinations thereof. The memory 114 may store data 116 and instructions 118 that are executed by the processor 112 to cause the user computing device 102 to perform operations.

[0041] In some implementations, the user computing device 102 may store or include one or more augmented reality models 120. For example, the augmented reality models 120 may be or otherwise include various machine learning models, such as neural networks (e.g., deep neural networks) or other types of machine learning models including nonlinear and / or linear models. The neural networks may include feedforward neural networks, recurrent neural networks (e.g., long short-term memory recurrent neural networks), convolutional neural networks, or other forms of neural networks. Exemplary augmented reality rendering models 120 are described with reference to FIGS. 5-9.

[0042] In some implementations, one or more augmented reality rendering models 120 may be received from a server computing system 130 over a network 180, stored in a user computing device memory 114, and then used or implemented by one or more processors 112. In some implementations, a user computing device 102 may implement multiple parallel instances of a single augmented reality rendering model 120.

[0043] More specifically, the augmented reality rendering model can utilize a perception model and a rendering model to render the augmented reality rendering in the template image. The perception model can be a model stored on the platform that is applicable with various rendering models. The rendering model can be generated by a third party using a software development kit. The rendering model can be sent to the platform after being built by the third party. In some implementations, the platform can receive data assets for the rendering model from the third party.

[0044] The template image may be processed by an augmented reality rendering model along with a perceptual model that generates a mesh and a segmentation mask based on processing of the template image, and the rendering model may process the template image, the mesh, and the segmentation mask to generate a pre-rendered image.

[0045] Additionally or alternatively, one or more augmented reality rendering models 140 may be included in or stored and implemented by a server computing system 130 that communicates with the user computing device 102 according to a client-server relationship. For example, the augmented reality rendering models 140 may be implemented by the server computing system 130 as part of a web service (e.g., a "pre-rendered try-on" service). Thus, one or more models 120 may be stored and implemented on the user computing device 102 and / or one or more models 140 may be stored and implemented on the server computing system 130.

[0046] The user computing device 102 may also include one or more user input components 122 that receive user input. For example, the user input component 122 may be a touch-sensitive component (e.g., a touch-sensitive display screen or touchpad) that senses the touch of a user input object (e.g., a finger or stylus). The touch-sensitive component functions to implement a virtual keyboard. Other exemplary user input components include a microphone, a traditional keyboard, or other means by which a user can provide user input.

[0047] The server computing system 130 includes one or more processors 132 and a memory 134. The one or more processors 132 may be any suitable processing device (e.g., a processor core, a microprocessor, an ASIC, an FPGA, a controller, a microcontroller, etc.) and may be a single processor or multiple processors operably connected. The memory 134 may include one or more non-transitory computer-readable storage media, such as RAM, ROM, EEPROM, EPROM, flash memory devices, magnetic disks, etc., and combinations thereof. The memory 134 may store data 136 and instructions 138 that are executed by the processor 132 to cause the server computing system 130 to perform operations.

[0048] In some implementations, the server computing system 130 includes or is implemented by one or more server computing devices. When the server computing system 130 includes multiple server computing devices, such server computing devices may operate according to a sequential computing architecture, a parallel computing architecture, or some combination thereof.

[0049] As described above, the server computing system 130 may store or include one or more machine-learned augmented reality rendering models 140. For example, the models 140 may be or otherwise include various machine-learned models. Exemplary machine-learned models include neural networks or other multi-layer nonlinear models. Examples of neural networks include feed-forward neural networks, deep neural networks, recurrent neural networks, convolutional neural networks, etc. Exemplary models 140 are described with reference to FIGS. 5-9.

[0050] The user computing device 102 and / or the server computing system 130 can train the models 120 and / or 140 through interaction with a training computing system 150 that is communicatively coupled via a network 180. The training computing system 150 may be separate from the server computing system 130 or may be part of the server computing system 130.

[0051] The training computing system 150 includes one or more processors 152 and memory 154. The one or more processors 152 may be any suitable processing device (e.g., a processor core, a microprocessor, an ASIC, an FPGA, a controller, a microcontroller, etc.) and may be a single processor or multiple processors operably connected. The memory 154 may include one or more non-transitory computer-readable storage media, such as RAM, ROM, EEPROM, EPROM, flash memory devices, magnetic disks, etc., and combinations thereof. The memory 154 may store data 156 and instructions 158 that are executed by the processor 152 to cause the training computing device 150 to perform operations. In some implementations, the training computing device 150 includes or is implemented by one or more server computing devices.

[0052] The training computing system 150 may include a model trainer 160 that trains the machine learning models 120 and / or 140 stored on the user computing device 102 and / or the server computing system 130 using various training or learning techniques, such as, for example, backpropagation of error. For example, a loss function may be backpropagated through the model to update one or more parameters of the model (e.g., based on the gradient of the loss function). Various loss functions may be used, such as mean squared error, likelihood loss, cross entropy loss, hinge loss, and / or various other loss functions. Gradient descent may be used to iteratively update the parameters over multiple training iterations.

[0053] In some implementations, performing backpropagation of the error may include performing truncated backpropagation over time. The model trainer 160 may perform a number of generalization techniques (e.g., weight decay, dropout, etc.) to improve the generalization ability of the model being trained.

[0054] In particular, model trainer 160 can train augmented reality models 120 and / or 140 based on a set of training data 162. Training data 162 can include, for example, shaders built by a third party using a software development kit, where the third party received the software development kit from the facilitating computing device or server computing system 130. The third party may have generated the shaders and data assets by building and testing an augmented reality experience using the software development kit.

[0055] In some implementations, if the user consents, the training examples can be provided by the user computing device 102. Thus, in such implementations, the model 120 provided to the user computing device 102 can be trained by the training computing system 150 based on user-specific data received from the user computing device 102. In some cases, this process can be referred to as personalizing the model.

[0056] In some implementations, the perception and rendering models of the augmented reality model may be trained using template images. The template images may be images from a corpus of template images that are provided to the user as pre-rendered images, or in some implementations, the training template images may be a separate set of template images used only for training.

[0057] Model trainer 160 includes computer logic utilized to provide desired functionality. Model trainer 160 can be implemented in hardware, firmware, and / or software controlling a general-purpose processor. For example, in some implementations, model trainer 160 includes program files stored in a storage device, loaded into memory, and executed by one or more processors. In other implementations, model trainer 160 includes one or more sets of computer-executable instructions stored in RAM, a hard disk, or a tangible computer-readable storage medium, such as an optical or magnetic medium.

[0058] Network 180 may be any type of communications network, such as a local area network (e.g., an intranet), a wide area network (e.g., the Internet), or some combination thereof, and may include any number of wired or wireless links. In general, communications on network 180 may be transmitted over any kind of wired and / or wireless connections using a variety of communications protocols (e.g., TCP / IP, HTTP, SMTP, FTP), encodings or formats (e.g., HTML, XML), and / or protection schemes (e.g., VPN, Secure HTTP, SSL).

[0059] The machine learning models described herein may be used in a variety of tasks, applications, and / or use cases.

[0060] In some implementations, an input to the machine learning model of the present disclosure may be image data. The machine learning model may process the image data to generate an output. As an example, the machine learning model may process the image data to generate an image recognition output (e.g., recognition of the image data, latent embedding of the image data, an encoded representation of the image data, a hash of the image data, etc.). As another example, the machine learning model may process the image data to generate an image segmentation output. As another example, the machine learning model may process the image data to generate an image classification output. As another example, the machine learning model may process the image data to generate an image data modification output (e.g., modification of the image data, etc.). As another example, the machine learning model may process the image data to generate an encoded image data output (e.g., an encoded and / or compressed representation of the image data, etc.). As another example, the machine learning model may process the image data to generate an upscaled image data output. As another example, the machine learning model may process the image data to generate a prediction output.

[0061] In some implementations, the input to the machine learning model of the present disclosure may be text or natural language data. The machine learning model may process the text or natural language data to generate an output. As an example, the machine learning model may process the natural language data to generate a language encoding output. As another example, the machine learning model may process the text or natural language data to generate a potential text embedding output. As another example, the machine learning model may process the text or natural language data to generate a translation output. As another example, the machine learning model may process the text or natural language data to generate a classification output. As another example, the machine learning model may process the text or natural language data to generate a text segmentation output. As another example, the machine learning model may process the text or natural language data to generate a semantic intent output. As another example, the machine learning model may process the text or natural language data to generate an upscaled text or natural language output (e.g., text or natural language data of higher quality than the input text or natural language). As another example, the machine learning model may process the text or natural language data to generate a prediction output.

[0062] In some implementations, the input to the machine learning model of the present disclosure may be latent encoding data (e.g., a latent spatial representation of the input, etc.). The machine learning model may process the latent encoding data to generate an output. As an example, the machine learning model may process the latent encoding data to generate a recognition output. As another example, the machine learning model may process the latent encoding data to generate a reconstruction output. As another example, the machine learning model may process the latent encoding data to generate a search output. As another example, the machine learning model may process the latent encoding data to generate a reclustering output. As another example, the machine learning model may process the latent encoding data to generate a prediction output.

[0063] In some implementations, the input to the machine learning model of the present disclosure may be sensor data. The machine learning model may process the sensor data to generate an output. As an example, the machine learning model may process the sensor data to generate a recognition output. As another example, the machine learning model may process the sensor data to generate a prediction output. As another example, the machine learning model may process the sensor data to generate a classification output. As another example, the machine learning model may process the sensor data to generate a segmentation output. As another example, the machine learning model may process the sensor data to generate a segmentation output. As another example, the machine learning model may process the sensor data to generate a visualization output. As another example, the machine learning model may process the sensor data to generate a diagnostic output. As another example, the machine learning model may process the sensor data to generate a detection output.

[0064] In some cases, the machine learning model may be configured to perform a task that includes encoding input data for reliable and / or efficient transmission or storage (and / or corresponding decoding). In another example, the input includes visual data (e.g., one or more images or videos) and the output comprises compressed visual data, and the task is a visual data compression task. In another example, the task may comprise generating an embedding of the input data (e.g., visual data).

[0065] In some cases, the input includes visual data and the task is a computer vision task. In some cases, the input includes pixel data of one or more images and the task is an image processing task. For example, the image processing task is image classification and the output is a set of scores, each score corresponding to a different object class and representing the likelihood that the one or more images depict an object belonging to that object class. The image processing task may be object detection and the image processing output identifies one or more regions in the one or more images and, for each region, the likelihood that the region represents an object of interest. As another example, the image processing task may be image segmentation and the image processing output defines, for each pixel in the one or more images, a respective likelihood for each category in a set of predefined categories. For example, the set of categories can be foreground and background. As another example, the set of categories can be object classes. As another example, the image processing task is depth estimation and the image processing output defines, for each pixel in the one or more images, a respective depth value. As another example, the image processing task is motion estimation and the network input includes multiple images and the image processing output defines, for each pixel in one of the input images, the motion of the scene depicted in the pixel between images in the network input.

[0066] 1A illustrates one exemplary computing system that can be used to implement the present disclosure. Other computing systems can be used as well. For example, in some implementations, the user computing device 102 can include a model trainer 160 and a training dataset 162. In such implementations, the model 120 can be both trained and used locally at the user computing device 102. In some such implementations, the user computing device 102 can implement the model trainer 160 to personalize the model 120 based on user-specific data.

[0067] 1B illustrates a block diagram of an exemplary computing device 10 for performing operations according to an exemplary embodiment of the present disclosure. The computing device 10 may be a user computing device or a server computing device.

[0068] The computing device 10 includes multiple applications (e.g., applications 1 through N). Each application includes its own machine learning library and machine learning model. For example, each application may include a machine learning model. Exemplary applications include a text messaging application, an email application, a dictation application, a virtual keyboard application, a browser application, etc.

[0069] 1B, each application may communicate with many other components of the computing device, such as, for example, one or more sensors, a context manager, a device state component, and / or additional components. In some implementations, each application may communicate with each device component using an API (e.g., a public API). In some implementations, the API used by each application is specific to that application.

[0070] 1C illustrates a block diagram of an exemplary computing device 50 for performing operations according to an exemplary embodiment of the present disclosure. The computing device 50 may be a user computing device or a server computing device.

[0071] Computing device 50 includes multiple applications (e.g., applications 1 through N). Each application communicates with a central intelligence layer. Exemplary applications include a text messaging application, an email application, a dictation application, a virtual keyboard application, a browser application, etc. In some implementations, each application can communicate with the central intelligence layer (and models stored therein) using an API (e.g., a common API across all applications).

[0072] The central intelligence layer includes multiple machine learning models. For example, as shown in FIG. 1C, each machine learning model (e.g., model) can be provided for each application and managed by the central intelligence layer. In other implementations, two or more applications can share a single machine learning model. For example, in some implementations, the central intelligence layer can provide a single model (e.g., a single model) for all applications. In some implementations, the central intelligence layer is included within or implemented by an operating system of the computing device 50.

[0073] The central intelligence layer can communicate with a central device data layer, which can be a central repository of data for the computing device 50. As shown in FIG. 1C, the central device data layer can communicate with many other components of the computing device, such as, for example, one or more sensors, a context manager, a device state component, and / or additional components. In some implementations, the central device data layer can communicate with each device component using an API (e.g., a private API).

[0074] Example Model Configuration 2 illustrates a block diagram of an example implementation 200 according to an example embodiment of the present disclosure. In some implementations, the example implementation 200 includes a search result list 204 describing results of a search query, and as a result of receiving the search result list 204, provides a pre-rendered image 208, which is a pre-rendered augmented reality image including a resulting product rendering of one or more results 206 from the search result list 204. Thus, in some implementations, the example implementation 200 can include a search query input area 202 operable to receive a search query input.

[0075] The exemplary implementation of FIG. 2 illustrates a search engine web service 210. A search engine input area 202 can capture a search query that can include one or more search terms. The platform can capture and process the search query to provide a search result list 204. The one or more search terms can be related to products for purchase. In these implementations, the search result list 204 can be one or more products related to the search terms. In some implementations, one or more results 206 from the search result list 204 can include one or more pre-rendered images 208 related to the result 206. In some implementations, the pre-rendered images 208 can include a rendering of the corresponding result 206. In some implementations, the user can further select a virtual live try-on augmented reality experience.

[0076] 3 illustrates a block diagram of an example implementation 300 according to an example embodiment of the present disclosure. The example implementation 300 is similar to the example implementation 200 of FIG. 2, except that the example implementation 300 further illustrates an example result.

[0077] FIG. 3 illustrates an example implementation 300 as a furniture try-on experience. For example, to generate the pre-rendered image shown in the implementation 300, a product 302 can be rendered in a template environment. In this example, a loveseat 304, a television 306, and a rug 308 are part of a template image of an example room. The template image can be one of multiple template images that can be processed to generate multiple pre-rendered images. In this example, the multiple pre-rendered images can include renderings of the product 302 in each corresponding template image. The template images can have various sizes, themes, and configurations. For example, the example pre-rendered image includes a television 306 across from a loveseat 304 with a rug 308 in between. In some implementations, the platform can allow the user to view the rendered product in various locations within the template image.

[0078] 4 illustrates a block diagram of an example implementation 400 according to an example embodiment of the present disclosure. The example implementation 400 is similar to the example implementation 200 of FIG. 2, except that the example implementation 400 further illustrates an example implementation in a mobile application.

[0079] In the exemplary implementation shown in FIG. 4, the platform is accessed through a user interface in a mobile application on a mobile device 402. A user can use the mobile device to access the mobile application, where the user can store preferences and access a library of pre-rendered images. For example, a user may prefer a particular living room template image 408 selected by the user. The user may have selected that template image 408 as the one that best matches the user's living room. The template image 408 may include a couch 404 and a similar lamp 406 as found in the user's home. The template image 408 may have been processed along with other template images to generate various pre-rendered images of various furniture and decor products. Using the mobile application, the platform can provide various decor or furniture products rendered in the user's preferred template image 408 to assist the user's own shopping experience. For example, a user may use the mobile application to see how a particular rug would look under a couch 404 and a lamp 406 similar to the setup in the user's living room.

[0080] 5 illustrates a block diagram of an example platform 500 according to an example embodiment of the present disclosure. In some implementations, the platform 500 is trained to receive requests and preferences 506 describing user-specific preferences, and to provide output data including pre-rendered images of augmented reality renderings on template images associated with the user-specific preferences 506 as a result of receiving the preferences 506. Thus, in some implementations, the platform 500 can include a directory 510 operable to store the template images, the augmented reality data assets, and the pre-rendered images.

[0081] The example platform of FIG. 5 includes a directory 510 of pre-rendered images, and a user 502. The user can have settings 504 selected by the user, which can include preferences 506. In some implementations, the preferences 506 can include a selected preference associated with a template image. The preferences 506 can be used to determine which pre-rendered image from a plurality of pre-rendered images can be provided to the user when a request is made. For example, a preference associated with a first template image can cause the platform to provide a pre-rendered image that includes a product rendered within the first template image.

[0082] The directory 510 can store template images, augmented reality assets, and pre-rendered images. The template images can be for various environments, including, but not limited to, rooms, gardens, driveways, and faces. The augmented reality assets can be used to render various objects and products, including, but not limited to, furniture, decorations, plants, electronics, automobiles, and makeup. In some implementations, the pre-rendered images can include a product rendered within the template image, in which case the product can be rendered based on the augmented reality assets.

[0083] In some implementations, the platform can store pre-rendered images of a variety of different products in a variety of environments. For example, dataset 1 can include various pre-rendered images 512 of lamps in various living rooms. Dataset 2 can include multiple pre-rendered images 514 of lipstick on a face. Dataset 3 can include multiple pre-rendered images 516 of trees in a garden. In some implementations, a user can use a user interface provided by a web service, mobile application, or kiosk to access the pre-rendered images to assist with shopping.

[0084] 9 illustrates a block diagram of an exemplary augmentation platform 900, according to an exemplary embodiment of the present disclosure. In some implementations, the augmentation platform 900 is trained to receive a set of input data describing a user request and provide output data including a pre-rendered rendering in a template image as a result of receiving the input data. Thus, in some implementations, the augmentation platform 900 can include an augmented reality pre-rendering platform 920 that is operable to interact with a user device and enable an augmented reality pre-rendering experience.

[0085] The augmented reality pre-rendering platform 920 shown in FIG. 9 includes a user interface 922 for enabling user interaction, a template library 924 for processing, a rendering engine 926 for processing the template images, and a pre-rendered library 928 that stores pre-rendered images generated by processing the template images.

[0086] In some implementations, the augmented reality pre-rendering platform 920 can receive user preferences 912 from the user computing device 910. The user preferences 912 can be used to determine which of the multiple pre-rendered images are provided to the user computing device 910 from the pre-rendered library 928.

[0087] In some implementations, the augmented reality pre-rendering platform 920 can provide an option for an augmented reality live try-on experience, which can involve processing of user media data or user camera feed 914 by a rendering engine 926 to generate renderings of user-provided data.

[0088] Exemplary Methods 6 illustrates a flow chart diagram of an exemplary method for performing in accordance with an exemplary embodiment of the present disclosure. Although FIG. 6 illustrates steps performed in a particular order for purposes of illustration and discussion, the method of the present disclosure is not limited to the particular order or arrangement illustrated. Various steps of method 600 may be omitted, rearranged, combined, and / or adapted in various ways without departing from the scope of the present disclosure.

[0089] At 602, the computing system can obtain a plurality of template images. In some implementations, the template images can include a room image, a face image, or a garden image. The plurality of template images can include images of a variety of environments. For example, the plurality of template images can include a variety of different room sizes, configurations, themes, lighting, or colors.

[0090] At 604, the computing system can process the plurality of template images with the augmented reality rendering model. The augmented reality rendering model can include object tracking and rendering. Processing of the template image with the augmented reality rendering model can generate a plurality of pre-rendered images. The plurality of pre-rendered images can include augmented reality renderings rendered within each of the plurality of template images. The augmented reality renderings can describe a product. The augmented reality rendering model can be based at least in part on data assets provided by a third party, where the data assets are generated by the third party by building an augmented reality rendering experience that renders the product in images and videos. For example, the product being rendered can be furniture (e.g., couch, chair, table, etc.). The plurality of pre-rendered images can include a particular couch rendered within a plurality of different template images depicting a variety of different rooms having different sizes, colors, themes, and configurations. In another example, the product can be a makeup product (e.g., lipstick, mascara, foundation, eyeliner, eyeshadow, etc.). In this implementation, the plurality of pre-rendered images can include a rendering of a makeup product such as lipstick. In this implementation, lipstick can be rendered onto multiple template face images.

[0091] In some implementations, processing with the augmented reality rendering model may include processing a plurality of template images with the perceptual model to generate a mesh and a segmentation mask, and then processing the mesh, the segmentation mask, and the template images with the rendering model to generate a plurality of pre-rendered images.

[0092] At 606, the computing system can receive a request and preferences for result images. The request can include search terms entered into a search engine. In some implementations, the preferences can include preselected templates.

[0093] At 608, the computing system can provide a pre-rendered result. The pre-rendered result can be based at least in part on the request and the preferences. The pre-rendered result can be a pre-rendered image from a plurality of pre-rendered images. In some implementations, the pre-rendered result can be a pre-rendered image that matches the preferences, where the preferences can include a template selected by the user from a plurality of template images.

[0094] In some implementations, the computing system may provide an augmented reality experience. The augmented reality experience may be provided upon receiving a selection to proceed to the augmented reality experience.

[0095] 7 shows a flow chart diagram of an exemplary method for performing in accordance with an exemplary embodiment of the present disclosure. Although FIG. 7 shows steps performed in a particular order for purposes of illustration and discussion, the method of the present disclosure is not limited to the particular order or arrangement shown. Various steps of method 700 may be omitted, rearranged, combined, and / or adapted in various ways without departing from the scope of the present disclosure.

[0096] At 702, a computing system can obtain an augmented reality asset. The augmented reality asset can include digitized parameters.

[0097] At 704, a computing system can obtain a plurality of template images.

[0098] At 706, the computing system can process the plurality of template images with the augmented reality model to generate a plurality of pre-rendered images. The plurality of pre-rendered images can be generated based at least in part on the digitization parameters. In some implementations, the computing system can generate a plurality of template models based at least in part on the plurality of template images. The augmented reality model can then process the plurality of template models to generate a plurality of pre-rendered images.

[0099] In some implementations, the computing system can include receiving an input to modify a template model of the plurality of template models. The template model can be modified based at least in part on the template. In some implementations, the template model can be provided for display.

[0100] The computing system may store the multiple pre-rendered images at 708. In some implementations, the multiple pre-rendered images may be stored on a server.

[0101] The computing system can provide the stored pre-rendered images to a user. In some implementations, the computing system can receive a search query including one or more search terms, the one or more search terms related to a product. The computing system can then provide search results, the search results including the pre-rendered image from the plurality of pre-rendered images. In some implementations, the pre-rendered image can include a rendering of the product.

[0102] In some implementations, the computing system may provide a link to a real-time augmented reality experience.

[0103] 8 shows a flow chart diagram of an exemplary method for performing in accordance with an exemplary embodiment of the present disclosure. Although FIG. 8 shows steps performed in a particular order for purposes of illustration and discussion, the method of the present disclosure is not limited to the particular order or arrangement shown. Various steps of method 800 may be omitted, rearranged, combined, and / or adapted in various ways without departing from the scope of the present disclosure.

[0104] At 802, a computing system can obtain an augmented reality asset. The augmented reality asset can include digitized parameters.

[0105] At 804, a computing system can obtain a plurality of template images.

[0106] At 806, the computing system can process the plurality of template images with the augmented reality model to generate a plurality of pre-rendered images. The augmented reality model can generate the plurality of pre-rendered images based at least in part on the digitization parameters.

[0107] At 808, the computing system can store the multiple pre-rendered images.

[0108] At 810, the computing system may receive a search query. The search query may include one or more search terms, where the one or more search terms are related to the product.

[0109] At 812, the computing system can provide search results. The search results can include pre-rendered images from a plurality of pre-rendered images retrieved from a server, and the pre-rendered images can include renderings of the products.

[0110] Additional Disclosures The technology described herein refers to servers, databases, software applications, and other computer-based systems, as well as actions taken on such systems and information sent to and from such systems. The inherent flexibility of computer-based systems allows for a wide variety of configurations, combinations, and divisions of tasks and functions among components. For example, the processes described herein may be implemented using a single device or component, or multiple devices or components operating in combination. Databases and applications may be implemented in a single system or distributed across multiple systems. Distributed components may operate sequentially or in parallel.

[0111] Although the subject matter of the present invention has been described in detail with respect to various specific exemplary embodiments thereof, each example is provided for illustrative purposes and not to limit the present disclosure. Those skilled in the art, upon understanding the foregoing, can easily create modifications, variations, and equivalents to such embodiments. Thus, the present disclosure is not intended to preclude inclusion of such modifications, variations, and / or additions to the subject matter as would be readily apparent to those skilled in the art. For example, features illustrated or described as part of one embodiment can be used with another embodiment to obtain yet another embodiment. Thus, the present disclosure is intended to cover such modifications, variations, and equivalents. [Explanation of symbols]

[0112] 50 Computing Devices 102 User Computing Devices 112 processors 114 Memory 116 Data 118 Command 120 Augmented Reality Rendering Models 120 Machine Learning Models 122 User Input Components 130 Server Computing System 132 processors 134 Memory 136 Data 138 Command 140 Augmented Reality Rendering Models 150 Training Computing System 152 processors 154 Memory 156 Data 158 Command 160 Model Trainer 162 training data 162 training dataset 180 Network 200 Implementation Example 202 Search query input area, search engine input area 204 Search Results List 206 results 208 Pre-rendered Images 210 Search Engine Web Services 300 Implementation Example 302 products 304 Loveseat 306 Television 308 Rug 400 Implementation Example 402 Mobile Devices 404 Couch 406 Lamp 408 Living Room Template Images 500 Platform 502 users 504 Configuration 506 Preference 510 Directory 512 Various pre-rendered images of lamps in various living rooms 514 Multiple pre-rendered images of lipstick on face 516 Multiple pre-rendered images of garden trees 600 ways 700 methods 800 ways 900 Extended Platform 910 User Computing Device 912 User Preferences 914 User Camera Feeds 920 Augmented Reality Pre-rendering Platform 922 User Interface 924 Template Library 926 Rendering Engine 928 Pre-rendered Library

Claims

1. 1. A computer-implemented method for providing a pre-rendered augmented image, comprising: obtaining, by a computing device, a plurality of template images; processing, by the computing device, the plurality of template images with an augmented reality rendering model to generate a plurality of pre-rendered images, the plurality of pre-rendered images including a product rendered within each of the plurality of template images; receiving, by the computing device, a request and preferences for a result image; providing, by the computing device, a pre-rendered result based at least in part on the request and the preferences, the pre-rendered result being a pre-rendered image from the plurality of pre-rendered images; A computer-implemented method comprising:

2. The computer-implemented method of claim 1 , wherein the plurality of template images comprises a plurality of room images.

3. The computer-implemented method of claim 1 , wherein the plurality of pre-rendered images comprises a plurality of rendered furniture images.

4. The computer-implemented method of claim 1 , wherein the plurality of template images comprises a plurality of facial images.

5. The computer-implemented method of claim 1 , wherein the plurality of pre-rendered images comprises a plurality of rendered makeup images.

6. The computer-implemented method of claim 1 , wherein the augmented reality rendering model comprises a perception model and a rendering model.

7. The computer-implemented method of claim 1 , wherein the request comprises a search term entered into a search engine.

8. The computer-implemented method of claim 1 , wherein the preferences comprise a preselected template.

9. 2. The computer-implemented method of claim 1, wherein the pre-rendered result is a particular pre-rendered image that matches the preferences, the preferences comprising a template selected by a user from the plurality of template images.

10. The computer-implemented method of claim 1 , further comprising providing an augmented reality rendering experience by the computing device.

11. 1. A computing system comprising: one or more processors; one or more computer-readable storage media that collectively store instructions that, when executed by the one or more processors, cause the computing system to perform operations; and wherein the operation comprises: obtaining an augmented reality asset, the augmented reality asset comprising digitized parameters; Obtaining a plurality of template images; processing the plurality of template images with the augmented reality model to generate a plurality of pre-rendered images based at least in part on the digitization parameters, the plurality of pre-rendered images including a product rendered within each of the plurality of template images; storing the plurality of pre-rendered images; receiving a request and preferences for a result image; providing a pre-rendered result based at least in part on the request and the preferences, the pre-rendered result being a pre-rendered image from the plurality of pre-rendered images; A computing system comprising:

12. The computing system of claim 11 , wherein storing the plurality of pre-rendered images comprises storing the plurality of pre-rendered images on a server.

13. receiving a search query comprising one or more search terms, the one or more search terms relating to a product; providing a search result, the search result comprising the pre-rendered image from the plurality of pre-rendered images, the pre-rendered image comprising a rendering of the product; and The computing system of claim 11 further comprising:

14. The computing system of claim 11 , wherein the augmented reality assets comprise data describing a product.

15. One or more computer-readable storage media that collectively store instructions that, when executed by one or more processors, cause a computing system to perform operations, the operations including: obtaining an augmented reality asset, the augmented reality asset comprising digitized parameters; Obtaining a plurality of template images; processing the plurality of template images with the augmented reality model to generate a plurality of pre-rendered images based at least in part on the digitization parameters, the plurality of pre-rendered images including a product rendered within each of the plurality of template images; storing the plurality of pre-rendered images on a server; receiving a search query comprising one or more search terms, the one or more search terms relating to a product; providing a search result, the search result comprising a pre-rendered image from the plurality of pre-rendered images retrieved from the server, the pre-rendered image comprising a rendered image of the product; One or more computer-readable storage media comprising:

Images