Imitating physical subjects in photos and videos with augmented reality virtual objects

摘要

Technologies for rendering augmented reality virtual objects include an augmented reality device to capture an image and recognize subjects within the image. Subjects may include persons or objects, such as accessories. The augmented reality device determines a context associated with the subjects of the image and selects one or more virtual objects based on the context. The identified context may include an activity or theme associated with the subject or accessories included in the image. Virtual objects may include accessories and clothing for virtual characters. The augmented reality device may select virtual objects that are similar to identified subjects and accessories, or that are associated with an identified activity or theme of the subject. The augmented reality device applies the virtual objects to a virtual character and renders an augmented reality scene based on the captured image that includes the virtual character. Other embodiments are described and claimed.

主权项

1. An augmented reality device for rendering augmented reality content, the augmented reality device comprising:
image capture circuitry to capture an image; subject recognition circuitry to (i) analyze the image to recognize a subject represented in the image and a background item in the image, (ii) access a character profile based on the subject, (iii) determine, based on the character profile, whether the background item is associated with the subject, (iv) determine, based on the character profile and in response to a determination that the background item is associated with the subject, a theme of the image, and (v) determine a context associated with the subject based on the image and the background item, wherein the context is related to the theme, wherein to determine the context comprises to classify a type of an accessory object represented in the image that is related to the subject; virtual object matching circuitry to select a virtual object based on the context of the subject and the background item, wherein the virtual object comprises an accessory object that is composable with the virtual character and is of the same type as the accessory object represented in the image, and wherein to select the virtual object comprises to select the virtual object based on similarity between the virtual object and a feature of the subject; and augmented reality rendering circuitry to (i) apply the virtual object to a virtual character and (ii) render an augmented reality scene based on the captured image and including the virtual character with the virtual object applied thereto.