Back to results
US 12,608,879 B2Granted
Generation of a 360-degree object view by leveraging available images on an online platform
AssigneeAdobe Inc.
InventorsGourav Singhal, Tridib Das, Sourabh Gupta
Life of this patent
Estimating the term…
Abstract
In some embodiments, a computing system generates a 360-degree view of a target object based on available images from an online platform. The computing system identifies a plurality of images having the same target object from one or more image sources on the online platform. The computing system categorizes the plurality of images into multiple view categories. The computing system then determines a representative image for each view category and generates a processed object image of the target object from the representative image for each view category. The computing system then stitches multiple processed object images to create a 360-degree view of the target object.
Claims19
A method in which one or more processing devices perform operations comprising:
- identifying, by an object detection module of a 360-degree view generation system, a plurality of images depicting a same target object from one or more image sources on an online platform, wherein the one or more image sources comprises user generated content on the online platform;
- categorizing, by an image categorization module of the 360-degree view generation system, the plurality of images into a plurality of view categories;
- determining, by an image aesthetics predictor module of the 360-degree view generation system, a representative image for each view category of the plurality of view categories;
- extracting, by a mask processing module of the 360-degree view generation system, an object image of the target object from the representative image for each view category to obtain a plurality of object images of the target object for the plurality of view categories;
- processing, by the mask processing module, the plurality of object images of the target object for the plurality of view categories to generate a plurality of processed object images, wherein said processing comprises:
- scaling the plurality of object images for the plurality of view categories to a predetermined scale to create a plurality of scaled object images with the same scale;
- and
- adjusting a plurality of attributes of the plurality of scaled object images to respective predetermined values to create a plurality of scaled and enhanced object images with the same attribute values corresponding to the plurality of attributes;
- and
- stitching, by an image stitching module of the 360-degree view generation system, the plurality of scaled and enhanced object images to create a 360-degree view of the target object.
7 more elements
A system, comprising:
- an object detection module configured to identify a plurality of images depicting a same target object from one or more image sources on an online platform, wherein the one or more image sources comprises user generated content on the online platform;
- an image categorization module configured to categorize the plurality of images into a plurality of view categories;
- an image aesthetics predictor module configured to determine a representative image for each view category of the plurality of view categories;
- a mask processing module configured to generate a plurality of processed object images of the target object from representative images for the plurality of view categories, wherein generating a plurality of processed object images of the target object from representative images for the plurality of view categories comprises:
- extracting an object image of the target object from the representative image for each view category to obtain a plurality of object images of the target object for the plurality of view categories;
- scaling the plurality of object images for the plurality of view categories to a predetermined scale to create a plurality of scaled object images with the same scale;
- and
- adjusting a plurality of attributes of the plurality of scaled object images to respective predetermined values to create a plurality of scaled and enhanced object images with the same attribute values corresponding to the plurality of attributes;
- an image stitching module configured to stitch the plurality of scaled and enhanced object images to create a 360-degree view of the target object;
- and
- a data store configured to store the 360-degree view of the target object,
- wherein the 360-degree view of the target object is periodically updated based on updates from the one or more image sources on the online platform.
9 more elements
A non-transitory computer-readable medium storing executable instructions, which when executed by a processing device, cause the processing device to perform operations comprising:
- identifying a plurality of images depicting a same target object from one or more image sources on an online platform, wherein the one or more image sources comprises user generated content on the online platform;
- a step for categorizing the plurality of images into a plurality of view categories;
- determining a representative image for each view category of the plurality of view categories;
- generating a plurality of scaled and enhanced object images, comprising:
- extracting an object image of the target object from the representative image for each view category to obtain a plurality of object images of the target object for the plurality of view categories,
- scaling the plurality of object images for the plurality of view categories to a predetermined scale to create a plurality of scaled object images with the same scale, and
- adjusting a plurality of attributes of the plurality of scaled object images to respective predetermined values to create a plurality of scaled and enhanced object images with the same attribute values corresponding to the plurality of attributes;
- and
- stitching the plurality of scaled and enhanced object images to create a 360-degree view of the target object.
6 more elements
Dependencies are inferred from the claim text.