Browse Prior Art Database

Environment Style Selector, Identifier, and Visualizer for Online Recommendations Disclosure Number: IPCOM000243605D
Publication Date: 2015-Oct-05
Document File: 2 page(s) / 110K

Publishing Venue

The Prior Art Database


Disclosed are a method and system for gathering user data and then using it to select from among a set of possible living environment recommendations for furniture, appliance, and/or decorative products via a commerce site.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 46% of the total text.

Page 01 of 2

Environment Style Selector, Identifier, and Visualizer for Online Recommendations

The presented solution improves on prior art in the area of environment personalization and recommendations. The novel contribution is a method and system for gathering user data and then using it to select from among a set of possible recommendations for products in a commerce site.

The system comprises a wearable device, camera, or other sensors to capture environment details. The system detects colors and materials commonly found and used in a user's home environment. The system then correlates these with the user's preferences, other known user preferences, and taste-based algorithms to identify likely matches for an empty space. The device uses these to recommend the best option out of potential decorative recommendations for the user. The system may further suggest

where a piece of decor or appliance usually found in similar environments might be

missing from the user's environment. The system can use visualization capabilities to display the selected item in the suggested location

The system integrates with a wearable sensor (e.g., two-dimensional/three-dimensional (2D/3D) camera) and performs the following steps:

1. Harvests video or still image data of the environment. The system harvests data from the camera. This information may be passively and continually harvested, passively harvested on a time interval basis (e.g., taking one shot every minute), or only harvested as part of an analysis of the background when the user is actively using the camera for other purposes (e.g., participating in a video chat). This may be controlled by user preferences and privacy settings. Some devices may provide subsidized modes which require the use of the novel system or require the use in a specific mode (e.g., similar to the subsidized electronic reader devices which display advertisements).

2. Analyzes the video for use of materials and colors. The system generates material and color signatures from the video. This may be done in the device, or by sending the video or still images to a cloud service. Boundary detection techniques are used to isolate materials and boundaries between objects.

Color detection is done trivially using existing techniques, though the system goes further and recognizes color combinations found in the user's environment, and the specific location of colors in the user's environment and to which materials they are applied. For instance, a user may have a white or off-white

color on the walls of a room, but has selected stainless steel for all consumer appliances and devices.

Each section of similar color is analyzed for material by comparing the variations in luminance, saturation, brightness, and other qualities in the area of the color. Object recognition is done using existing machine learning techniques to assign the materials and colors to objects. For instance, the novel system recognizes stainless steel appliances, whit...