Super-personalized game item

Abstract

The two-dimensional image is converted (302) into a 3D neural radiation field (NeRF), which is modified (402) based on text personalized according to the player and the input, to resemble the character's equipment required by the text (1700). A model scores (404) the degree of matching of an image with a row of text to produce a final 3D NeRF that can be converted (408) into a polygonal mesh and imported into a computer simulation, such as a computer game.

Description

Technical Field

[0001] The present application generally relates to the rapid generation of hyper-personalized game items. Background Art

[0002] As understood herein, creating characters (such as non-player characters (NPCs)) and their equipment for computer simulations (such as computer games) can be very time-consuming. Summary of the Invention

[0003] As further understood herein, it may be desirable for game developers to be able to create characters and equipment for video games in a simple, fast, and intuitive manner.

[0004] Accordingly, the apparatus includes at least one computer storage device, the at least one computer storage device not being a transient signal, and the computer storage device further includes instructions that are executable by at least one processor to generate a base mesh or a Neural Radiance Field (NeRF) from a plurality of images. It should be noted that text-to-2D image (e.g., Stable Diffusion) or text-to-mesh techniques can be used directly, but in all embodiments, the technique is applied to the real-time creation of hyper-personalized in-game items.

[0005] In the context of NeRF, the instructions are executable to use the text input into a Contrastive Language-Image Pretraining (CLIP) model to generate a modified NeRF from the base NeRF and convert the modified NeRF into a polygonal mesh representing the virtual character equipment for rendering the equipment in at least one computer simulation.

[0006] In some examples, the CLIP model rates the degree of match between an image and text.

[0007] Preferably, the text is derived from player information, such as the title of at least one computer simulation that the player is interested in. The text can describe the equipment of a game character, such as a mask. The instructions can be executable to generate text from a starting phrase using learned subsequent phrases.

[0008] In another aspect, the device includes at least one processor programmed with instructions to receive a text description of equipment personalized for player data. At least partially based on the text description, the instructions are executable to generate a virtual three-dimensional (3D) equipment within less than two minutes after receiving the text description and render the virtual equipment on a display.

[0009] In another aspect, a method includes: receiving text based on data related to a player of a computer simulation, and generating a Neural Radiance Field based on the text starting from a base model.

[0010] The details of both the structure and operation of the present application can be best understood with reference to the accompanying drawings, in which like reference numerals refer to like components, and in the drawings: Description of the Drawings

[0011] Figure 1 is a block diagram of an example system according to this principle;

[0012] Figure 2 illustrates example initial logic in an example flowchart format;

[0013] Figure 3 illustrates creating a 3D neural radiance field (NeRF) representing a human head from 2D images;

[0014] Figure 4 illustrates example text-based NeRF customization logic in an example flowchart format;

[0015] Figure 4A is another representation of the example logic;

[0016] Figures 5 to 8 illustrates a customized NeRF using a first descriptive phrase;

[0017] Figure 9 illustrates a customized NeRF using a second descriptive phrase;

[0018] Figure 10 illustrates a customized NeRF using a third descriptive phrase;

[0019] Figure 11 illustrates a series of image generation steps for learning how to customize a general NeRF;

[0020] Figure 12 illustrates a data structure for generating text;

[0021] Figure 13 illustrates additional text-based example NeRF customization logic in an example flowchart format;

[0022] Figure 14 illustrates text-based example equipment customization logic in an example flowchart format;

[0023] Figure 15 illustrates in connection with Figure 14 a consistent example screenshot;

[0024] Figure 16 illustrates in connection with Figure 14 a consistent example player-customized game equipment; and

[0025] Figure 17 and Figure 18Shows an example screenshot of providing customized equipment for a specific player. Detailed implementation

[0026] The present disclosure generally relates to computer ecosystems, which include aspects of a network of consumer electronics (CE) devices, such as, but not limited to, computer game networks. The systems herein may include server and client components that can be connected via a network, such that data can be exchanged between the client and the server components. The client components may include one or more computing devices, including game consoles (such as, Sony or game consoles manufactured by Microsoft or Nintendo or other manufacturers), extended reality (XR) headsets (such as, virtual reality (VR) headsets, augmented reality (AR) headsets), portable TVs (e.g., smart TVs, Internet-enabled TVs), portable computers (such as laptops and tablets), and other mobile devices (including smartphones and additional examples discussed below). These client devices can operate in a variety of operating environments. For example, some of the client computers may employ, for example, the Linux operating system, Microsoft's operating system, or the Unix operating system, or the operating systems produced by Apple Inc. or Google, or the Berkeley Software Distribution or Berkeley Standard Distribution (BSD) operating system (including descendants of BSD). These operating environments can be used to execute one or more browsing programs, such as browsers made by Microsoft or Google or Mozilla or other browser programs that can access websites hosted by the Internet servers discussed below. Moreover, the operating environments according to the present principles can be used to execute one or more computer game programs.

[0027] A server and / or gateway can be used, which may include one or more processors that execute instructions that configure the server to receive and transmit data via a network such as the Internet. Alternatively, the client and the server can be connected via a local intranet or a virtual private network. The server or the controller can be instantiated by a game console (such as, Sony Play )), a personal computer, etc.

[0028] Information can be exchanged between the client and the server via the network. For this purpose, and for security reasons, the server and / or the client may include a firewall, a load balancer, a temporary storage device, a proxy, and other network infrastructures to ensure reliability and security. One or more servers can form a device that implements a method of providing a secure community (such as an online social website or a game player network) to network members.

[0029] The processor can be a single-chip or multi-chip processor, and the processor can perform logic by means of various lines such as address lines, data lines, and control lines, as well as registers and shift registers. A processor including a digital signal processor (DSP) can be an implementation of the circuit.

[0030] The components included in one implementation can be used in other implementations in any suitable combination. For example, any one of the various components described herein and / or depicted in the drawings can be combined, interchanged, or excluded from other implementations.

[0031] A "system having at least one of A, B, and C" (similarly, "a system having at least one of A, B, or C" and "a system having at least one of A, B, and C") includes a system having only A, only B, only C, A and B together, A and C together, B and C together, and / or A, B, and C together.

[0032] Now referring to Figure 1 , an example system 10 is shown, and the example system can include one or more of the example devices mentioned above and further described below in accordance with the principles of the present invention. The first of the example devices included in system 10 is a consumer electronics (CE) device, such as an audio-video device (AVD) 12, such as but not limited to a theater display system that can be based on a projector, or an Internet-enabled TV with a TV tuner (equivalent to a set-top box for controlling the TV). Alternatively, the AVD 12 can also be a computerized Internet-enabled ("smart") phone, tablet, laptop, head-mounted device (HIVID), and / or headphones (such as smart glasses or VR headphones), another wearable computerized device, a computerized Internet-enabled music player, a computerized Internet-enabled headphone, a computerized Internet-enabled implantable device (such as an implantable skin device), etc. In any case, it should be understood that the AVD 12 is configured to adopt these principles (for example, communicate with other CE devices to adopt these principles, perform the logic described herein, and perform any other functions and / or operations described herein).

[0033] Therefore, to adopt such principles, the AVD 12 can be established by some or all of the components shown. For example, the AVD 12 can include one or more touch displays 14, which can be implemented by a high-definition or ultra-high-definition "4K" or higher flat screen. The touch display 14 can include, for example, a capacitive or resistive touch sensing layer having an electrode grid for touch sensing consistent with these principles.

[0034] The AVD 12 may also include: one or more speakers 16 for outputting audio according to the present principles; and at least one additional input device 18, such as an audio receiver / microphone, for inputting audible commands to the AVD 12 to control the AVD 12. The exemplary AVD 12 may also include one or more network interfaces 20 for communicating under the control of one or more processors 24 over at least one network 22 (such as the Internet, WAN, LAN, etc.). Thus, the interface 20 may be, but is not limited to, a Wi-Fi transceiver, which is an example of a wireless computer network interface, such as, but not limited to, a mesh network transceiver. It should be understood that the processor 24 controls the AVD 12 to employ the present principles, including other elements of the AVD 12 described herein, such as controlling the display 14 to present an image thereon and receive input therefrom. Additionally, note that the network interface 20 may be a wired or wireless modem or router, or other suitable interface, such as a wireless telephone transceiver or a Wi-Fi transceiver as mentioned above, etc.

[0035] In addition to the above, the AVD 12 may also include one or more input and / or output ports 26, such as a high-definition multimedia interface (HDMI) port or a universal serial bus (USB) port, for physically connecting to another CE device and / or a headphone port for connecting headphones to the AVD 12 to present audio from the AVD 12 to the user via the headphones. For example, the input port 26 may be connected to a cable or satellite source 26a of audio-video content via a wired or wireless manner. Thus, the source 26a may be a separate or integrated set-top box, or a satellite receiver. Or the source 26a may be a game console or a disk player containing content. When the source 26a is implemented as a game console, it may include some or all of the components described below with respect to the CE device 48.

[0036] The AVD 12 may also include one or more computer memories / computer-readable storage media 28, such as disk-based or solid-state storage devices that are not transient signals, which in some cases are embodied as separate devices or as part of a personal video recording device (PVR) or video disk player within the chassis of the AVD, and these devices are located inside or outside the AVD chassis for playing back AV programs, or as removable storage media or the servers described below. Moreover, in some embodiments, the AVD 12 may include a positioning or location receiver, such as, but not limited to, a cell phone receiver, a GPS receiver, and / or an altimeter 30, which is configured to receive geolocation information from a satellite or a cell phone base station and provide the information to the processor 24 and / or determine the altitude at which the AVD 12 is located together with the processor 24.

[0037] Continuing the description of the AVD 12, in some embodiments, the AVD 12 may include one or more cameras 32, which may be thermal imaging cameras, digital cameras (such as webcams), infrared sensors, event-based sensors, and / or cameras integrally formed into the AVD 12 and controllable by the processor 24 to collect pictures / images and / or videos in accordance with this principle. The AVD 12 may also include a transceiver 34 and other near field communication (NFC) elements 36 for communicating with other devices using Bluetooth and / or NFC technologies, respectively. Example NFC elements may be radio frequency identification (RFID) elements.

[0038] In addition, the AVD 12 may include one or more auxiliary sensors 38 that provide input to the processor 24. For example, one or more of the auxiliary sensors 38 may include one or more pressure sensors that form a layer of the touch display 14 itself and may be, but are not limited to, piezoelectric pressure sensors, capacitive pressure sensors, piezoresistive strain gauges, optical pressure sensors, electromagnetic pressure sensors, etc. Other sensor examples include pressure sensors, motion sensors (such as accelerometers, gyroscopes, odometers, or magnetic sensors), infrared (IR) sensors, optical sensors, speed and / or rhythm sensors, event-based sensors, gesture sensors (e.g., for sensing gesture commands). Thus, the sensor 38 may be implemented by one or more motion sensors, such as individual accelerometers, gyroscopes, and magnetometers, and / or an inertial measurement unit (IMU), typically including a combination of an accelerometer, a gyroscope, and a magnetometer, to determine the position and orientation of the AVD 12 in three dimensions, or by an event-based sensor (such as an event detection sensor (EDS)). The output provided by an EDS consistent with this disclosure indicates a change in the light intensity sensed by at least one pixel of a light sensing array. For example, if the light sensed by a pixel is decreasing, the output of the EDS may be -1; if the sensed light is increasing, the output of the EDS may be +1. An output binary signal of 0 may indicate no change in light intensity below a certain threshold.

[0039] The AVD 12 may also include a wireless TV broadcast port 40 for receiving OTA TV broadcasts that provide input to the processor 24. In addition to the above, it should be noted that the AVD 12 may also include an infrared (IR) transmitter and / or IR receiver and / or IR transceiver 42, such as an IR Data Association (IRDA) device. A battery (not shown) may be provided to power the AVD 12, and a kinetic energy harvester may also be provided to convert kinetic energy into electrical energy to charge the battery and / or power the AVD 12. A graphics processing unit (GPU) 44 and a field programmable gate array 46 may also be included. One or more haptic / vibration generators 47 may be provided for generating haptic signals that can be sensed by a person holding or touching the device. Thus, the haptic generator 47 may use an electric motor to vibrate all or a part of the AVD 12, where the electric motor is connected via a rotatable shaft of the motor to an eccentric and / or unbalanced weight such that the shaft can be rotated under the control of the motor (which in turn can be controlled by a processor, such as the processor 24) to create vibrations of various frequencies and / or amplitudes and force simulations in various directions.

[0040] A light source, such as a projector, such as an infrared (IR) projector, may also be included.

[0041] In addition to the AVD 12, the system 10 may also include one or more other CE device types. In one example, the first CE device 48 may be a computer game console that can be used to send computer game audio and video to the AVD 12 via commands sent directly to the AVD 12 and / or via a server as described below, while the second CE device 50 may include components similar to those of the first CE device 48. In the example shown, the second CE device 50 may be configured as a computer game controller manipulated by a player or a head-mounted display (HMD) worn by the player. The HMD may include a head-up transparent or opaque display for presenting AR / MR content or VR content (more generally, extended reality (XR) content) respectively. The HMD may be configured as a glasses-type display or as a larger VR-type display sold by a computer game device manufacturer.

[0042] In the example shown, only two CE devices are shown, but it should be understood that fewer or more devices may be used. The devices herein may implement some or all of the components shown in the AVD 12. Any of the components shown in the following figures may be incorporated into some or all of the components shown in the case of the AVD 12.

[0043] Now referring to at least one of the aforementioned servers 52, which includes at least one server processor 54, at least one tangible computer-readable storage medium 56 (such as a disk-based or solid-state storage device), and at least one network interface 58. The network interface, under the control of the server processor 54, allows communication with the other illustrated devices via network 22 and can in fact facilitate communication between the server and client devices in accordance with this principle. It should be noted that the network interface 58 can be, for example, a wired or wireless modem or router, a Wi-Fi transceiver, or other suitable interface (such as a wireless telephone transceiver).

[0044] Thus, in some embodiments, the server 52 can be an Internet server or an entire server "farm" and can include and execute "cloud" functions such that the devices of system 10 can access a "cloud" environment via the server 52 in an example embodiment related to, for example, a network game application. Alternatively, the server 52 can be implemented by one or more game consoles or other computers located in the same room or nearby as the other illustrated devices.

[0045] The components shown in the following figures may include some or all of the components shown herein. Any user interface (UI) described herein can be combined and / or extended, and UI elements can be mixed and matched between UIs.

[0046] This principle may employ various machine learning models, including deep learning models. Machine learning models consistent with this principle can use various algorithms trained in various ways, including supervised learning, unsupervised learning, semi-supervised learning, reinforcement learning, feature learning, self-learning, and other forms of learning. Examples of such algorithms can be implemented by computer circuits including one or more neural networks, such as convolutional neural networks (CNNs), recurrent neural networks (RNNs), and an RNN called long short-term memory (LSTM) networks. Support vector machines (SVMs) and Bayesian networks can also be considered examples of machine learning models. In addition to the network types described above, the models herein can also be implemented by classifiers.

[0047] As understood herein, performing machine learning can thus involve accessing training data and then training the model so that the model can process more data for inference. Thus, an artificial neural network / artificial intelligence model trained by machine learning may include an input layer, an output layer, and a plurality of hidden layers therebetween, the plurality of hidden layers being configured and weighted to infer an appropriate output.

[0048] Now referring to Figure 2 and Figure 3。The initial logic executable by any one or more processors herein begins at block 200 by receiving a set of head images (labeled 300 in Figure 3 ), which may be 2D images of the same underlying head generated by one or more artists from various different perspectives. Based on these images, a 3D base head ( Figure 3 labeled 302 in ) is generated at block 202. The 3D base head 302 may be a 3D Neural Radiance Field (NeRF), which can be considered a 3D volume stored in a machine learning (ML) model. In a specific example, the base head 302 may be a NeRF that can be quickly trained (within two minutes and in some examples, within one minute) after inputting the following text description to produce a modified NeRF discussed further below. To this end, an example NeRF is a NeRF encoded in a multiresolution hash table, which serves to store the positions of a sparse multiresolution 3D grid to speed up the training and rendering of the NeRF.

[0049] Note that in some embodiments, a head scan of a real-world head may be used to generate the Figure 3 base head 302 in, rather than an image.

[0050] Now turning to Figure 4 . Once the base head 302 has been generated, a text description of the modified base head can be received at block 400 of the desired head. The text can be input from a text input device (such as a keyboard), and / or from a speech-to-text conversion of an oral description, and / or from an initial starting phrase followed by learned additional descriptive phrases as discussed further herein. Alternatively, more primitive algorithmic techniques can be used to generate text in a madlib style such that answers to a survey can be inserted into the text describing the target head.

[0051] Moving to block 402, the text is input into a NeRF modification engine, which may include a fully connected (non-convolutional) deep network. The input to the modification engine can be a single continuous 5D input including values representing three spatial dimensions and two view dimensions, and the output of the engine can include the volume density and view-dependent emitted radiance at spatial positions represented by associated 3D values.

[0052] Proceed to block 404, where the output of the NeRF modification engine (which can be regarded as the modified NeRF) is evaluated to determine how well the output matches the input text description. In one example, this evaluation can be performed by a Contrastive Language-Image Pretraining (CLIP) model, which rates how well an image matches an input text string. CLIP is an open-source model that scores how well a line of text matches an image. The CLIP model can be trained using cosine similarity on image-text pairs (such as photos from the Internet with captions) to score the goodness of match. Thus, the modification of the base NeRF is essentially controlled using a loss function that depends on the text input to the CLIP model.

[0053] Decision diamond 406 indicates that if it is determined after the current modification iteration that the loss has not reached the threshold (target small) loss, then at block 408 the current iteration of the modified NeRF is modified again, and the logic loops back to block 404. However, once it is determined that the loss is small enough, the logic moves to block 410 to output the final modified NeRF. It should be noted that a certain number of loops (e.g., one hundred) can be executed instead of a loss threshold. Blocks 402 through 408 can use gradient descent techniques.

[0054] In some examples, the logic can then move to block 412 to convert the modified NeRF into a polygon mesh, which can be imported into a computer simulation (such as a computer game) at block 414 and rendered on a display at block 416 during gameplay. Block 412 can also generate materials for the mesh, such as albedo, roughness, and normal textures. The modified NeRF can be converted into a polygon mesh with the same topology as the modified NeRF. This can be done by using the marching cubes technique. Alternatively, the modified NeRF can be converted into a polygon mesh by substantially shrink-wrapping the modified NeRF by placing a mesh on the modified NERF and recording which parts of the mesh touch the surface of the modified NeRF when the mesh is simulated as being evacuated.

[0055] Figure 4A is an alternative illustration. Starting from block 420, a NeRF is generated according to parameters that may include values determining the appearance of the head and is tuned to make the final NeRF look like the input image. Moving to block 422, the NeRF can be rendered from multiple camera angles and enhancements (such as color, rotation, and affine) can be applied at block 424.

[0056] Continuing to block 426, the CLIP image embedding may be retrieved and used in block 428 to compare the current iteration of NeRF to the text image embedding to obtain a loss value. At block 430, the loss value is added to other terms, such as the "blurriness" of NeRF, by obtaining the total variance of the depth of the rendered image. If it is desired to modify the parameters at block 420, the total loss is back-propagated at block 432 using gradient descent.

[0057] Figures 5 to 8 shows the input text "Old tough guy's photorealistic avatar" from Figure 3 Four corresponding perspective views of the modified NeRF 500 generated from the base NeRF 302 in FIG. The modified NeRF 500 is generated in less than two minutes after the text is input into the model.

[0058] Figure 9 and Figure 10 Shown are modified NeRF 900, NeRF 1000, respectively, which are generated from base NeRF 302 using corresponding input texts "photorealistic headshot of sea monster" and "photorealistic headshot of devil", respectively.

[0059] Figure 11

[0066] The ML models described above (particularly those for mesh techniques) can be trained based on a chain of causal relationships from initial parameters (variables) controlling the vertices of an object to the pixels rendered on the screen. Starting from vertex 1100, the chain of causality can continue to primitives 1102 based on vertex 1100, and then further to fragments 1104 where the primitives are at least partially filled (e.g., with textures). From fragments 1104, processed fragments 1106, to which color can be added, can be generated and merged into pixel outputs 1108 rendered on the screen. Arrow 1110 shows that this chain can be traversed back down to the original vertex 1100 to see how specific values ​​in the chain affect the final pixel image 1108, which can be used to train the ML model.

[0060] It should be noted that if a mesh of the head is used instead of a radiance field, the ML model can perform differentiable rendering, where parameters can be varied to see in real time how the changes affect the output image. When the new face begins to match the target defined in the text, the ML model uses a loss function to minimize the loss.

[0061] As mentioned above, learned subsequent phrases may be used to generate part or all of the input text describing the desired modified NeRF from the starting phrase. Figure 12This is shown. The prompt template 1200 can be presented on the developer's computer screen, where there are multiple phrases to fill in the player / developer's answers. Additional text 1202 can be presented for the answers to each prompt. The additional text 1202 can be generated by an ML model trained to complete sentences based on the input phrases.

[0062] Figure 13 Additional techniques are shown. Starting from block 1300, text indicating a character biography and the equipment the character uses to start the game is received. The text-to-image generator 1302 receives the text and generates an image consistent with the principles herein, which is imported into the computer game at block 1304 and presented when the game is played at block 1306. It should be noted that block 1302 can alternatively use a text-to-3D mesh generator (such as a NeRF generator) or an image (such as a decal on the player's equipment), but the more primary use case is to generate actual 3D assets in the form of a mesh / material.

[0063] It can now be understood that the 3D NeRF images modified from the text are generated in real time, rather than drawn, Photoshop-processed, found, or preconceived. Instead, they are generated based on other text-based image generation.

[0064] Therefore, the present principle provides a technique for generating a coherent 3D head from text within a few minutes (e.g., two minutes or less or one minute or less). Starting from a base model, new NeRFs are generated using text in a specific domain. The use of the hash table encoding of the NeRFs helps with quick editing (~1 minute).

[0065] Now turn to Figures 14 to 18 , where the use of other base models besides the head is shown. For example, the base models of in-game items such as hats or weapons can be used to produce highly personalized game items.

[0066] Starting from Figure 14 at block 1400, player data of a specific individual player is obtained. Preferably, the data is non-sensitive player data, such as first-party games played previously, to generate unique in-game items.

[0067] Move to block 1402 and create text based on the data from block 1100. For example, if the player data indicates that the player is a fan of Game X, the system creates a text prompt that includes or refers to "Game X", which is used at block 1404 in accordance with the present principles to generate an image of an item of equipment, such as different in-game masks, which can be presented on the player's display at block 1406 and, if desired, on the player's character during gameplay at block 1408, for example, based on the player's acceptance. The image can include materials such as texture data.

[0068] The text generator can be rather primitive. For example, Figure 15 The final text prompt 1500 shown in (and presented on the screen together with the subsequent image 1600 in Figure 16 ) is "Kabuki mask" and can include "Based on [Character Y] from [Game X]". A generic prompt can be "Mask of [Style] based on [Character] from [Game] ([Release Year])".

[0069] Figure 17 and Figure 18 are further shown. A user interface (UI) 1700 can be presented on the player's display 1702, prompting (1704) the player as to whether they wish to customize an item of equipment for the player's character (PC), in this case a mask. One or more selectors 1706 can be presented, allowing the player to accept or not. If the player accepts, a text input can be automatically generated in accordance with the above principles to generate an image of the item of equipment, which can be presented to the player for viewing at 1800 in Figure 18 .

[0070] Figure 18 It is further shown that selectors 1802 can be presented to allow the player to select certain options, including the option to purchase the customized equipment so that the same equipment is not available to other players. The player may also wish to share the equipment with other players.

[0071] The above tools and techniques can be provided in a final user game computing device, such as a computer game console, such that final user game players can use the tools described herein in a game (i.e., as part of playing a computer game) to create and / or modify game objects for each other.

[0072] Although specific embodiments have been shown and described in detail herein, it should be understood that the subject matter covered by the present invention is limited only by the claims.

Claims

1. A device, which comprises: at least one computer storage device, which is not a transient signal and comprises instructions executable by at least one processor to: generate a neural radiance field (NeRF) from a plurality of images; input text into a contrastive language-image pre-training (CLIP) model to generate a modified NeRF from the base NeRF; and convert the modified NeRF into a polygonal mesh representing virtual character equipment for rendering the equipment in at least one computer simulation.

2. The device according to claim 1, wherein the CLIP model rates the degree of match between the image and the text.

3. The device according to claim 2, wherein the text is derived from player information.

4. The device according to claim 3, wherein the player information comprises the title of at least one computer simulation.

5. The device according to claim 1, wherein the text describes character equipment.

6. The device according to claim 5, wherein the equipment comprises a mask.

7. The device according to claim 1, wherein the instructions are executable to: generate the text from a starting phrase using learned subsequent phrases.

8. The device according to claim 1, which comprises the at least one processor.

9. An apparatus, which comprises: at least one processor programmed with instructions to: receive a text description of equipment personalized according to player data; generate a virtual three-dimensional (3D) equipment within less than two minutes after receiving the text description at least in part based on the text description; and render the virtual equipment on a display.

10. The apparatus according to claim 9, wherein the instructions are executable to: generate the virtual equipment within less than one minute after receiving the text description.

11. The apparatus according to claim 9, wherein the virtual equipment comprises a modified neural radiance field (NeRF).

12. The apparatus according to claim 11, wherein the modified NeRF comprises a modified NeRF containing a hash table.

13. The apparatus according to claim 11, wherein the instructions are executable to: input text into a contrastive language-image pre-training (CLIP) model to generate the modified NeRF from a base NeRF; and convert the modified NeRF into a polygonal mesh representing virtual equipment for rendering the virtual equipment in at least one computer simulation.

14. The apparatus according to claim 13, wherein the CLIP model rates the degree of match between the image and the text.

15. The apparatus according to claim 9, wherein the instructions are executable to: generate the virtual equipment using a machine learning (ML) model by minimizing a loss indicator matching the descriptive text.

16. The apparatus according to claim 15, wherein the ML model comprises at least one fully connected deep network.

17. The apparatus according to claim 15, wherein the input to the ML model comprises values representing three spatial dimensions and two view dimensions, and the output of the ML model comprises volumetric density and view-dependent emitted radiance.

18. A method, which comprises: receiving text based on data related to a computer-simulated player; and generating a neural radiance field based on the text starting from a base model.

Images