US20250316034
2025-10-09
Physics
G06T19/006
The system leverages augmented reality (AR) and virtual reality (VR) technologies to create virtual representations of physical spaces, such as homes or offices. Users can interact with these virtual environments using voice commands to preview virtual objects, like furniture, within these spaces. The system generates these environments from images uploaded by users, enabling them to visualize how products will look in their actual settings before making a purchase.
A commerce system or mobile application generates virtualized environments based on source images of physical spaces. Users can navigate these environments using a voice-based interface. By speaking the name of a room or object, the system retrieves and displays virtual representations of the specified items. This process is facilitated by natural-language processing (NLP) functions that interpret voice commands into actionable instructions for navigating and modifying the virtual environment.
The system employs scanning functions to capture images of physical spaces using various optical sensors like cameras and LIDAR. These images are used to build a coordinate plane of the region, allowing for accurate placement of virtual objects. Additionally, the system creates an object table that lists items within each region, helping users locate and interact with objects effectively.
Users can manually input identifiers for specific rooms or rely on machine-learning operations to automatically recognize and label regions based on common features. When users issue voice commands, the system references these identifiers to display the appropriate virtual environment. The app allows users to manipulate virtual objects through touch-based inputs or predefined configurations, maintaining contextual awareness to avoid unrealistic object collisions.
The client app can recognize new objects from images captured in physical stores, generating virtual representations for these items within the user's specified regions. Users can provide verbal instructions to preview these objects in their desired locations. The app situates objects based on user inputs or preconfigured defaults, ensuring realistic placement relative to other items and avoiding overlaps that defy real-world physics.