Google Patent | Shadows For Inserted Content
Patent: Shadows For Inserted Content
Publication Number: 20200273238
Publication Date: 20200827
Applicants: Google
Abstract
Systems and methods for generating shadows for inserted content are provided. The inserted content may include augmented reality content that is inserted into an image of a physical space. An example includes determining a location to insert content within an image. The content may include a polygonal mesh defined in part by a skeleton that has a plurality of joints. Examples may further include selecting a plurality of selected joints form the plurality of joints. Examples may also include generating a shadow polygon based on the content and determining shadow contributions values for the plurality of selected joints for pixels of the shadow polygon. Examples may also include combining the shadow contribution values from the selected joints to generate shadow magnitude values for the pixels, rendering the shadow polygon using the shadow magnitude values, and overlaying the inserted content on the rendered shadow polygon.
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is a divisional of, and claims priority to, U.S. patent application Ser. No. 16/150,813, filed on Oct. 3, 2018, entitled “Shadows for Inserted Content”, which claims priority to U.S. Provisional Patent Application No. 62/568,083, filed on Oct. 4, 2017 and claims priority to U.S. Provisional Patent Application No. 62/568,115, filed on Oct. 4, 2017, the disclosures of which are incorporated herein by reference in their entireties.
[0002] This application is related to U.S. application Ser. No. 16/150,794, filed on Oct. 3, 2018, entitled “Shadows for Inserted Content”, now U.S. Pat. No. 10,607,403, which is incorporated herein by reference in its entirety.
BACKGROUND
[0003] Content may be inserted into an image or a user’s field of view. For example, an augmented reality (AR) system may generate an immersive augmented environment for a user by inserting content. The immersive augmented environment can be generated by superimposing computer-generated content on a user’s field of view of the real world. For example, the computer-generated content can include labels, textual information, images, sprites, and three-dimensional entities. These images may be displayed at a position in the user’s field of view so as to appear to overlay an object in the real world. Similarly, the computer-generated content may be overlaid on a displayed image. The inserted content may generate shadows that overlay the displayed image. Existing technology for generating shadows may be inadequate for use in real-time AR applications.
SUMMARY
[0004] This disclosure describes systems and methods for generating shadows for inserted content. For example, the inserted content may include augmented reality content that is inserted into an image of a physical space.
[0005] One aspect is a method comprising: determining a location to insert content within an image, the content including a polygonal mesh defined in part by a skeleton that has a plurality of joints; selecting a plurality of selected joints form the plurality of joints; generating a shadow polygon; determining shadow contributions values for the plurality of selected joints for pixels of the shadow polygon (e.g., a shadow contribution value for each of the plurality of selected joints); combining the shadow contribution values from the selected joints to generate shadow magnitude values for the pixels; rendering the shadow polygon using the shadow magnitude values; and overlaying the inserted content on the rendered shadow polygon.
[0006] Another aspect is a non-transitory computer-readable storage medium comprising instructions stored thereon that, when executed by at least one processor, are configured to cause a computing system to at least: capture an image with a camera assembly; identify a surface plane within the image; determine a location to insert content, the content including a polygonal mesh defined in part by a skeleton that has a plurality of joints; select a plurality of selected joints form the plurality of joints; assign fade-out values to the plurality of selected joints; generate a shadow polygon based on the content; determine shadow contributions values for the plurality of selected joints for pixels of the shadow polygon; combine the shadow contribution values from the selected joints to generate shadow magnitude values for the pixels; normalize the shadow magnitude values to generate normalized shadow magnitude values; apply a radial falloff to the normalized shadow magnitude values to generated adjusted shadow magnitude values; render the shadow polygon using the adjusted shadow magnitude values; and overlay the inserted content on the rendered shadow polygon.
[0007] Yet another aspect is a system comprising: at least one processor; and memory storing instructions that, when executed by the at least one processor, cause the system to: identify a surface plane within the image; determine a location to insert content, the content including a skeletal animation model; generate a bounding box on the surface plane for the content based on a plurality of skeletal joints from the skeletal animation model; determine a center of mass location of the content based on projecting the plurality of skeletal joints on the surface plane and averaging positions of the projected skeletal joints; generate a first shadow polygon on the surface plane based on the bounding box and the center of mass location, the first shadow polygon having an oval shape; render the first shadow polygon based on applying a radial falloff function to the center of mass location to generate a rendered first shadow polygon; select a plurality of selected skeletal joints form the plurality of skeletal joints; generate a second shadow polygon based on the content; determine shadow contributions values for the plurality of selected skeletal joints for pixels of the second shadow polygon; combine the shadow contribution values from the selected skeletal joints to generate shadow magnitude values for the pixels of the second shadow polygon; render the second shadow polygon using the shadow magnitude values to generate a rendered second shadow polygon; combine the rendered first shadow polygon and the rendered second shadow polygon to generate a combined shadow polygon; and overlay the inserted content on the combined shadow polygon.
[0008] Another aspect is a method comprising: determining a location to insert content within an image, the content including a polygonal mesh defined in part by a skeleton that has a plurality of joints; generating a bounding box for the content based on at least some of the plurality of joints; determining a center of mass location of the content based on at least some of the plurality of joints; generating a first shadow polygon based on the bounding box and the center of mass location; generating first shadow magnitude values for pixels of the first shadow polygon based on the center of mass location; selecting a plurality of selected joints form the plurality of joints; generating a second shadow polygon based on the content; determining shadow contributions values for the plurality of selected joints for pixels of the second shadow polygon; combining the shadow contribution values from the selected joints to generate second shadow magnitude values for the pixels of the second shadow polygon; combining the first shadow magnitude values from the first shadow polygon and the second shadow magnitude values from the second shadow polygon to generate a combined shadow magnitude values; and rendering a combined shadow polygon using the combined shadow magnitude values.
[0009] The details of one or more implementations are set forth in the accompanying drawings and the description below. Other features will be apparent from the description and drawings, and from the claims.
BRIEF DESCRIPTION OF THE DRAWINGS
[0010] FIG. 1 is a block diagram illustrating a system according to an example implementation.
[0011] FIG. 2 is a third person view of an example physical space, in which a user is experiencing an AR environment through the example HMD of FIG. 1.
[0012] FIGS. 3A, 3B, and 3C are diagrams depicting an example head-mounted display device and controller, in accordance with implementations as described herein.
[0013] FIG. 4 is a schematic view of a user experiencing the AR environment via an example portable electronic device.
[0014] FIG. 5 is a diagram of an example method of generating shadows for inserted content, in accordance with implementations described herein.
[0015] FIG. 6 is a diagram of an example method of generating shadows for inserted content, in accordance with implementations described herein.
[0016] FIGS. 7A-7H are schematic diagrams of steps of generating shadows for inserted content in accordance with implementations as described herein.
[0017] FIG. 8 is a diagram of an example method of generating shadows for inserted content, in accordance with implementations described herein.
[0018] FIGS. 9A-9I are schematic diagrams of steps of generating shadows for inserted content in accordance with implementations as described herein.
[0019] FIG. 10 shows an example of a computer device and a mobile computer device that can be used to implement the techniques described herein.
DETAILED DESCRIPTION
[0020] Reference will now be made in detail to non-limiting examples of this disclosure, examples of which are illustrated in the accompanying drawings. The examples are described below by referring to the drawings, wherein like reference numerals refer to like elements. When like reference numerals are shown, corresponding description(s) are not repeated and the interested reader is referred to the previously discussed figure(s) for a description of the like element(s).
[0021] Augmented reality (AR) systems include systems that insert computer-generated content into a user’s perception of the physical space surrounding the user. The computer-generated content may include labels, textual information, images, sprites, and three-dimensional entities. In some implementations, the content is inserted for entertainment, educational, or informational purposes.
[0022] An example AR system is a portable electronic device, such as a smartphone, that includes a camera and a display device. The portable electronic device may capture images using the camera and show AR images on the display device that include computer-generated content overlaid upon the images captured by the camera.
[0023] Another example AR system includes a head-mounted display (HMD) that is worn by a user. The HMD includes a display device that is positioned in front of a user’s eyes. For example, the HMD may occlude the user’s entire field of view so that the user can only see the content displayed by the display device. In some examples, the display device is configured to display two different images, one that is viewable by each of the user’s eyes. For example, at least some of the content in one of the images may be slightly offset relative to the same content in the other image so as to generate the perception of a three-dimensional scene due to parallax. In some implementations, the HMD includes a chamber in which a portable electronic device, such as a smartphone, may be placed so as to permit viewing of the display device of the portable electronic device through the HMD.
[0024] Another example AR system includes a HMD that permits the user to see the physical space while the HMD is being worn. The HMD may include a micro-display device that displays computer-generated content that is overlaid on the user’s field of view. For example, the HMD may include an at least partially transparent visor that includes a combiner that permits light from the physical space to reach the user’s eye while also reflecting images displayed by the micro-display device toward the user’s eye.
[0025] When computer-generated content is inserted into an image, shadows may be generated around or below the content so that the content appears more realistic. For example, a three-dimensional model may be rendered using soft shadows that appear to have been generated by a large-area overhead light source. These soft shadows may be preferable to hard shadows that may be generated point/directional lights because information about the location of point/directional lights may not be available for the physical space.
[0026] AR systems may need to refresh images displayed to a user in real time at a high rate, such as 24 frames per second (FPS), 30 FPS, 60 FPS, or another rate. Traditional techniques for generating shadows may require determining or estimating lighting in the physical space. But determining or estimating lighting in a scene from an image may require so many computations (or processor cycles) that it cannot be performed in real-time on an AR system at an acceptable frame rate. Some traditional techniques require prior information about the lighting in an environment, which is may not be available for many of the environments in which an AR system is used. The techniques described further herein allow for generating shadows for inserted content in a realistic manner without requiring prior knowledge of the environmental lighting while using fewer processor cycles than traditional techniques. Additionally, due to the reduced number of processing cycles required by the techniques described herein, these techniques may allow for inserting content into a captured image/video in a realistic manner while using less power than traditional techniques would require. This reduction in power required to estimate lighting and provide shadows for inserted content may be particularly important in AR systems that include battery-operated mobile devices.
[0027] An example AR system captures images of the physical space surrounding a user. The system may then identify a surface plane, such as the ground or a floor, in the image and determine a location to insert content. For example, the system may receive a user input indicating a location on the screen for the content. The content may be placed at the location indicated by the user or at a location on the identified surface plane that is below the location indicated by the user. The content may, for example, include a three-dimensional model that is animated using a skeletal animation model. A skeletal animation model may include a mesh and a set of connected skeletal joints (which may be referred to as a skeleton or a rig) that is used to animate and position the mesh. The skeletal joints may be represented as three-dimensional coordinates. In some implementations, the three-dimensional coordinates are defined with respect to a common origin of the skeletal animation model. The skeletal animation model may also store connection data that define segments that connect the joints. These segments may be analogous to bones of skeleton. The segments connecting the joints may move or rotate about at least some of the joints. These movements may result in corresponding changes in the outer surface mesh of the skeletal animation model. As the segments move or rotate, connected joints and segments may also move or rotate. In some implementations, the joints (e.g., skeletal joints) can be an approximation of joints of a skeletal animation model. In some implementations, one or more joints can be at, or can include, an intersection of longitudinal members of content (e.g., an object). In some implementations, a skeleton can be, or can be referred to as a frame.
[0028] Next, the system may generate a bounding box and a shadow center point (e.g., a shadow middle point) on the surface plane based on the content. For example, the bounding box may be a rectangular shape on the surface planes that circumscribes all of the joints of a skeletal animation model associated with the content (or a projection of the joints onto the plane). The shadow center point may be a center of mass of the joints. For example, the center of mass may be calculated by averaging the positions of the joints (or the positions of the joints after they have been projected onto the surface plane). In some implementations, the center of mass can be an approximate center of mass. The joints may be weighted equally or may be weighted based on other factors such as distance from the surface plane. In some implementations, not all of the joints are used to generate a bounding box and shadow center point. For example, in some implementations, the inserted content may identify a subset of joints that are to be used in generating the bounding box and shadow center point (i.e., a joint whitelist). In some implementations, the inserted content may identify a subset of joints that are to be excluded when generating the bounding box and shadow center point (i.e., a joint blacklist). For example, the inserted content may be associated with a data structure that includes a joint blacklist or joint whitelist. Each of the skeletal joints of the inserted content may include a Boolean value that indicates whether the joint should be used to generate shadows (e.g., in generating the bounding box and shadow center point).
[0029] Based on the bounding box and the shadow center point, a shadow polygon may be generated on the surface plane. The shadow polygon may have various shapes. For example, the shadow polygon may have an oval shape that fits within the bounding box and includes a first and second axis that intersect at the shadow center point. Although much of the description is related to an oval shape, the shape can be any type of oblong shape.
[0030] During rendering, the shadow polygon may be shaded with a transparency value (e.g., an alpha value) that increases with distance from the shadow center point (i.e., the shadow polygon becomes more transparent and, therefore, less visible further from the shadow center point). In some examples, the transparency value increases non-linearly based on distance from the shadow center point. The shadow polygon can then be placed behind the content to be inserted as a first shadow entity.
[0031] Additionally or alternatively, a second shadow entity can be generated for the content to be inserted. The first shadow entity, which is described above, may generate a single shadow that has a radial falloff from the determined location of the shadow center point. This second shadow entity may have a non-uniform shape based on the joints of the skeletal animation model. Some implementations include the first shadow entity, some implementations include the second shadow entity, and some implementations include both the first shadow entity and the second shadow entity.
[0032] To generate the second shadow entity, some implementations generate a second shadow polygon on the surface plane below the location of the content to be inserted. For example, the second shadow polygon may be a rectangle having the same shape as a bounding box of the joints of the skeletal animation model (or a selected subset of joints). The second shadow polygon may comprise a polygon with sixteen or another number of sides, such as eight, ten, twelve, fourteen, fifteen, seventeen, or eighteen sides. Other numbers of sides are possible as well. In some implementations, the sides of the polygon approximate an oval that fits in the bounding box.
[0033] Next, a selection of the content may be identified. In some implementations, the selection includes all of the content or a portion of the content. For example, the selection of the content may comprise 25% of the joints in the skeletal animation model associated with the content. The 25% may be selected as the lowest 25% of the joints (i.e., the 25% of the joints having the lowest positional value along a vertical dimension) or as the 25% of the joints that are closest to the surface plane. In some implementations, a different threshold value is used to select joints instead of 25%. For example, in some implementations, the threshold may be 10%, 20%, 30%, 33%, 40%, 50%, or another value. The selection may also be based on a number of joints to be selected, such as 10, 20, 50, 100, or another number of joints. In some implementations, the joints in the selection may be assigned a fade-out factor. The fade-out factor is higher for joints that were close to not being selected. For example, if 25% of joints are selected based on distance to the surface plane, the selected joints that are furthest from the surface plane may have the highest fade-out value. The fade-out value may limit some of the joints’ contribution to a shadow so as to prevent popping artifacts that may occur as joints move in and out of the selection during sequential frames. The selection of joints may also be based on values stored in a data structure associated with the inserted content. For example, the data structure may identify some joints to exclude from use in generating shadows. In this case, a threshold percentage or number of joints may identified from the non-excluded joints.
[0034] The joints may be assigned a radius value, which may be used to calculate a shadow contribution for the joint. In some implementations, all of the selected joints are assigned the same radius value. Alternatively, the joints may have different radius values that correspond to properties of the model (e.g., the joints that are farthest from other joints may have a larger radius than other joints). For example, the joints having larger radius values may contribute more to a generated shadow. In some implementations, the radius values of some joints are set to zero so as to exclude those joints from contributing to shadows.
[0035] A shadow strength (also can be referred to as a shadow magnitude) may then be calculated for each pixel of the second shadow polygon. For example, the shadow strength may be calculated during rendering. The pixel value may be calculated by selecting a maximum shadow contribution from each of the joints in the selection. In other implementations, the pixel value is selected by averaging or summing the shadow contributions of the selected joints.
[0036] For each joint, the shadow contribution to a particular pixel may be based on multiple factors, such as a distance factor and an angle of elevation factor. For example, the shadow contribution may be determined by combining the distance factor and the angle of elevation factor. In some implementations, the distance factor corresponds to the geometric solid angle of the joint with respect to the pixel (i.e., corresponding to the space subtended by the joint). In some implementations, the solid angle is calculated as a value that is proportional to the arctangent of the quotient of the radius of the joint divided by the distance to the joint. In some implementations, the distance factor approximates the solid angle and is calculated as the quotient of the radius of the joint divided by the sum of the radius of the joint plus the distance to the joint from the pixel’s location in the scene. This approximation of the solid angle may require less processor cycles than calculating the solid angle using the arctangent function. In some implementations, the angle of elevation factor is based on the projection of the vector from the pixel to the joint and the normal vector of the second shadow polygon. For example, in some implementations, the angle of elevation factor is calculated using the following formula:
f*dot(N,V)*r/(r+d) where:
[0037] N is the surface normal of the surface plane;
[0038] V is the unit vector from the pixel to the joint;
[0039] r is the radius of the joint;
[0040] d is the distance from the pixel to the joint;* and*
[0041] f is the fade-out value for the joint.
[0042] Once the shadow strengths are calculated for each pixel, a gamma function may be applied to normalize the shadow strength. For example, the gamma function may remap the shadow strengths to normalized values that accentuate the midtones without having excessively dark regions. Additionally, a smooth radial falloff may be applied to the second shadow polygon to eliminate hard shadow edges at the polygon border. The smooth radial falloff may be applied in a manner similar to that described for the first shadow polygon.
[0043] The inserted content and one or more generated shadow entities may then be presented to the user (e.g., overlaid on a captured image of the physical space surrounding the user, projected/displayed on an optical combiner disposed within the user’s field of view, etc.). In some implementations, the first and second shadow entities described above are blended together or otherwise combined. For example, polygons corresponding to the first and second shadow entities may be combined during rendering each pixel by selecting and using the lower transparency value from the first or second shadow entity for each pixel or by combining the values in another way. Although many example herein refer to transparency values (or alpha values), other implementations are possible as well. For example, some implementations calculate a shadow strength rather than a transparency value. The shadow strength would be proportional to the opacity of the shadow entities that are described in terms of transparency/alpha values (e.g., the shadow strength would be highest when the transparency of the shadow entity/polygon is lowest, and the shadow strength would be lowest when the transparency of the shadow entity/polygon is highest). In these implementations, rather than overlaying a partially transparent dark colored polygon over the image, the shadow strength is used to alter the image. For example, the value of a pixel may be multiplied by one minus the shadow strength, wherein the shadow strength has a value between zero and one. In this manner, the pixel value gets darker as the shadow strength increases. In some implementations, a pixel value can include a color of a pixel.
[0044] Although many examples described herein relate to AR systems inserting visual content into an AR environment, content may be inserted using the techniques described herein in other systems too. For example, the techniques described herein may be used to insert content into an image or video.
[0045] FIG. 1 is a block diagram illustrating a system 100 according to an example implementation. The system 100 generates an augmented reality (AR) environment for a user of the system 100. In some implementations, the system 100 includes a computing device 102, a head-mounted display device (HMD) 104, and an AR content source 106. Also shown is a network 108 over which the computing device 102 may communicate with the AR content source 106.
[0046] The computing device 102 may include a memory 110, a processor assembly 112, a communication module 114, a sensor system 116, and a display device 118. The memory 110 may include an AR application 120, AR content 122, an image buffer 124, an image analyzer 126, a content analyzer 128, and a shadow engine 130. The computing device 102 may also include various user input components (not shown) such as a controller that communicates with the computing device 102 using a wireless communications protocol. In some implementations, the computing device 102 is a mobile device (e.g., a smart phone) which may be configured to provide or output AR content to a user via the HMD 104. For example, the computing device 102 and the HMD 104 may communicate via a wired connection (e.g., a Universal Serial Bus (USB) cable) or via a wireless communication protocol (e.g., any WiFi protocol, any BlueTooth protocol, Zigbee, etc.). In some implementations, the computing device 102 is a component of the HMD 104 and may be contained within a housing of the HMD 104.
[0047] The memory 110 can include one or more non-transitory computer-readable storage media. The memory 110 may store instructions and data that are usable to generate an AR environment for a user.
[0048] The processor assembly 112 includes one or more devices that are capable of executing instructions, such as instructions stored by the memory 110, to perform various tasks associated with generating an AR environment. For example, the processor assembly 112 may include a central processing unit (CPU) and/or a graphics processor unit (GPU). For example, if a GPU is present, some image/video rendering tasks, such as generating shadows or shading polygons representing shadows, may be offloaded from the CPU to the GPU.
[0049] The communication module 114 includes one or more devices for communicating with other computing devices, such as the AR content source 106. The communication module 114 may communicate via wireless or wired networks, such as the network 108.
[0050] The sensor system 116 may include various sensors, such as a camera assembly 132. Implementations of the sensor system 116 may also include other sensors, including, for example, an inertial motion unit (IMU) 134, a light sensor, an audio sensor, an image sensor, a distance and/or proximity sensor, a contact sensor such as a capacitive sensor, a timer, and/or other sensors and/or different combination(s) of sensors.
[0051] The IMU 134 detects motion, movement, and/or acceleration of the computing device 102 and/or the HMD 104. The IMU 134 may include various different types of sensors such as, for example, an accelerometer, a gyroscope, a magnetometer, and other such sensors. A position and orientation of the HMD 104 may be detected and tracked based on data provided by the sensors included in the IMU 134. The detected position and orientation of the HMD 104 may allow the system to detect and track the user’s gaze direction and head movement.
[0052] In some implementations, the AR application may use the sensor system 116 to determine a location and orientation of a user within a physical space and/or to recognize features or objects within the physical space.
[0053] The camera assembly 132 captures images and/or videos of the physical space around the computing device 102. The camera assembly 132 may include one or more cameras. The camera assembly 132 may also include an infrared camera.
[0054] The AR application 120 may present or provide the AR content to a user via the HMD and/or one or more output devices of the computing device 102 such as the display device 118, speakers, and/or other output devices. In some implementations, the AR application 120 includes instructions stored in the memory 110 that, when executed by the processor assembly 112, cause the processor assembly 112 to perform the operations described herein. For example, the AR application 120 may generate and present an AR environment to the user based on, for example, AR content, such as the AR content 122 and/or AR content received from the AR content source 106. The AR content 122 may include content such as images or videos that may be displayed on a portion of the user’s field of view in the HMD 104. The AR environment may also include at least a portion of the physical (real-world) environment and physical (real-world) entities. For example, shadows may be generated so that the content better fits the physical space in which the user is located. The content may include objects that overlay various portions of the physical space. The content may be rendered as flat images or as three-dimensional (3D) objects. The 3D objects may include one or more objects represented as polygonal meshes. The polygonal meshes may be associated with various surface textures, such as colors and images.
[0055] The AR application 120 may use the image buffer 124, image analyzer 126, content analyzer 128, and shadow engine 130 to generate images for display via the HMD 104 based on the AR content 122. For example, one or more images captured by the camera assembly 132 may be stored in the image buffer 124. In some implementations, the image buffer 124 is a region of the memory 110 that is configured to store one or more images. In some implementations, the computing device 102 stores images captured by the camera assembly 132 as a texture within the image buffer 124. Alternatively or additionally, the image buffer may also include a memory location that is integral with the processor assembly 112, such as dedicated random access memory (RAM) on a GPU.
[0056] The image analyzer 126 may determine various properties of the image, such as the location of a surface plane upon which the content may be positioned. In some implementations, the surface plane is a substantially horizontal plane that corresponds to the ground, a floor, or another surface upon which objects, such as the content to be inserted, could be placed.
[0057] The AR application 120 may determine a location to insert content. For example, the AR application may prompt a user to identify a location for inserting the content and may then receive a user input indicating a location on the screen for the content. The AR application may determine the location of the inserted content based on that user input. For example, the location for the content to be inserted may be the location indicated by the user. In some implementations, the location is determined by mapping the location indicated by the user to a plane corresponding to a surface such as a floor or the ground in the image (e.g., by finding a location on a plane identified by the image analyzer 126 that is below the location indicated by the user). The location may also be determined based on a location that was determined for the content in a previous image captured by the camera assembly (e.g., the AR application may cause the content to move across a surface that is identified within the physical space captured in the image).
[0058] The content analyzer 128 may then determine various properties of the content to be inserted at the determined location. For example, the content may be associated with a 3D model and skeletal animation model that includes joints. The skeletal animation model may be disposed within the 3D model and may allow for movement of portions of the 3D model around some or all of the joints. As an example, the content analyzer 128 may determine a bounding box and shadow center point on the surface plane based on the location of at least some of the joints of the skeletal animation model. For example, the skeletal joints may be projected onto the surface plane. In at least some embodiments, the joints are projected from an overhead position so as to generate shadows that appear to come from an overhead light source (e.g., by discarding the height component (i.e., the Y component when the surface is parallel to the X-Z plane) of the 3D position of the joints or setting the height component equal to the height of the plane). In some implementations, all of the joints are used to generate the bounding box and identify the shadow center point. In some implementations, a subset of the joints are used to generate the bounding box and identify the shadow center point (e.g., the inserted content may identify joints to use or exclude). In some implementations, the shadow center point may not be at a center of an object.
[0059] The bounding box may be a rectangle on the surface that contains all of the projected joints. In at least some implementations, the rectangle is aligned with the axes of the 3D coordinate system (e.g., if the surface is parallel to the X-Z plane, the sides of the rectangle are aligned with either the X or Z axes).
[0060] The shadow center point can be determined in various ways. For example, the shadow center point can be the spatial midpoint of the projected joints. The shadow center point can also be calculated as a center of mass of the projected joints (i.e., the average position of the projected joints). In some implementations, the joints may be assigned weights for purposes of calculating the center of mass. For example, the weights can be assigned based on distance from the surface (e.g., the joints that are closer to the surface have a higher weight than those that are further away).
[0061] The content analyzer 128 may also select a plurality of the joints to generate a plurality of selected joints. For example, the content analyzer 128 may select a predetermined percentage of the joints based on distance to the surface plane. In some implementations, the predetermined percentage is 25%, however, other predetermined percentages can be used too. Additionally or alternatively, a predetermined quantity of the joints can be selected. In some implementations, all of the joints are selected. A subset of joints may also be selected. In some implementations, a subset of the joints are selected based on a data structure associated with the inserted content. Beneficially, by selecting a subset of joints, the amount of processor cycles used to generate shadows may be reduced. The content analyzer 128 may also assign a fade-out value to the selected joints. For example, the fade-out value of a joint may be proportional to the distance between the joint and the surface plane. The content analyzer may also assign radius values to the selected joints. In some implementations, a same radius value is assigned to each of the selected joints. For example, the radius value may be determined based on the size of the content (e.g., the radius may be a predetermined percentage of the size of the content in one dimension, such as the longest dimension of the content). Additionally, different radius values may be assigned to the selected joints. In these implementations, the radius values may be based on distance from the selected joint to the next closest joint in the skeletal model.
[0062] The shadow engine 130 may generate one or more shadows for the content to be inserted. In some implementations, the shadow engine 130 generates a first shadow polygon based on the bounding box and shadow center point determined by the content analyzer 128. The first shadow polygon may have a dark color (e.g., black) and a transparency value that varies based on distance from the shadow center point. In some implementations, the transparency value is determined by applying a non-linear falloff based on distance from the center point. The non-linear falloff may cause the pixels near the center of the polygon to have a low transparency value and the pixels near the edges of the polygon to have a higher transparency value. In at least some implementations, the pixels on the edge of the polygon are completely transparent.
[0063] The shadow engine 130 may also generate a second shadow polygon that is shaded based, at least in part, on the selected joints. For example, each pixel of the second shadow polygon may be shaded according to a shadow strength that is determined based on various properties of the pixel with respect to the selected joints. For example, a shadow contribution value may be calculated with respect to a particular pixel of the second shadow polygon and a particular selected joint. The shadow contribution value may be based on a distance and an overhead angle (also known as an elevation angle) of the particular selected joint relative to the particular pixel (i.e., a selected joint that is closer and/or more directly overhead a pixel makes a stronger shadow contribution than a selected joint that is more distance and/or less directly overhead). Additionally, the shadow contribution value may be reduced (or otherwise scaled) based on the fade-out value assigned to the selected joint (i.e., the selected joints with higher fade-out values will have lower shadow contribution values, all other things being equal, than selected joints with lower fade-out values).
[0064] The shadow engine 130 may then combine the shadow contributions from the selected joints to generate a shadow strength value for a particular pixel of the second shadow polygon. For example, the shadow contribution values may be combined by selecting a maximum value from the shadow contributions. Alternatively, the shadow contribution values may be combined by summing or averaging at least some of the shadow contributions values.
[0065] The shadow engine 130 may then combine the first shadow polygon and the second shadow polygon. The transparency values discussed with respect to the first shadow polygon and the shadow strengths discussed with respect to the second shadow polygon are generally inverses of one another. For example, a transparency value may be subtracted from a maximum transparency value (e.g., 1.0) to generate a shadow strength value and vice versa. In some implementations, a first shadow strength from the first shadow polygon and a second shadow strength from the second shadow polygon are combined by taking the square root of the sum of the square of the first shadow strength and the square of the second shadow strength.
[0066] In some implementations, the shadow engine 130 may also use other techniques to generate shadows. For example, the shadow engine 130 may use shadow maps to generate shadows. An example technique for generating shadow maps is described in Williams, Lance. “Casting Curved Shadows on Curved Surfaces.” ACM Siggraph Computer Graphics, Vol. 12, No. 3, ACM, 1978. The shadows generated by the shadow maps may be combined with one or more of the first shadow polygon and the second shadow polygon. In some implementations, the shadows generated using shadow maps are combined with one or more of the first shadow polygon and the second shadow polygon by weighting the shadows based on distance to the inserted content. For example, the shadows generated using shadow maps may be weighted more heavily when the inserted content is closer to the camera assembly 132. The first shadow polygon or the second shadow polygon may be given greater weight when the inserted content is further away from the camera assembly 132 (e.g., to substitute for the shadows generated using shadow maps). Additionally, other techniques to generate shadows may be used too and combined (or weighted) in a similar manner. In some implementations, the shadowing techniques described herein can be combined with a shadow map to produce an overall more realistic overall shadow.
[0067] In some implementations, the image analyzer 126, content analyzer 128, and shadow engine 130 may include instructions stored in the memory 110 that, when executed by the processor assembly 112, cause the processor assembly 112 to perform operations described herein to generate an image or series images that are displayed to the user (e.g., via the HMD 104).
[0068] The AR application 120 may update the AR environment based on input received from the camera assembly 132, the IMU 134, and/or other components of the sensor system 116. For example, the IMU 134 may detect motion, movement, and/or acceleration of the computing device 102 and/or the HMD 104. The IMU 134 may include various different types of sensors such as, for example, an accelerometer, a gyroscope, a magnetometer, and other such sensors. A position and orientation of the HMD 104 may be detected and tracked based on data provided by the sensors included in the IMU 134. The detected position and orientation of the HMD 104 may allow the system to detect and track the user’s position and orientation within a physical space. Based on the detected position and orientation, the AR application 120 may update the AR environment to reflect a changed orientation and/or position of the user within the environment.
[0069] Although the computing device 102 and the HMD 104 are shown as separate devices in FIG. 1, in some implementations, the computing device 102 may include the HMD 104. In some implementations, the computing device 102 communicates with the HMD 104 via a cable, as shown in FIG. 1. For example, the computing device 102 may transmit video signals and/or audio signals to the HMD 104 for display for the user, and the HMD 104 may transmit motion, position, and/or orientation information to the computing device 102.
[0070] The AR content source 106 may generate and output AR content, which may be distributed or sent to one or more computing devices, such as the computing device 102, via the network 108. In an example implementation, the AR content includes three-dimensional scenes and/or images. Additionally, the AR content may include audio/video signals that are streamed or distributed to one or more computing devices. The AR content may also include an AR application that runs on the computing device 102 to generate 3D scenes, audio signals, and/or video signals.
……
……
……
