Glasses Shadow Rendering via Morphed Face Model

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current virtual try-on technologies for glasses lack realism due to inefficiencies in generating shadows, requiring detailed 3D models of faces and laborious ray tracing processes.

Innovation Solution

A method that uses a 3D generic face model morphed to match a user's face, combined with a transformed glasses model, and a lighting model to simulate shadows, which are then applied to images of the user's face using a shadow transform, allowing for efficient rendering of realistic shadows without the need for detailed user-specific models.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If ray tracing is used to generate shadows on a detailed 3D model of the user's face, then the realism of the virtual try-on is improved, but the time consumption and computational complexity increase significantly

Engineering Contradiction:

Improverealism of virtual try-onVSAvoidtime consumption for model generation and ray tracing

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent uses a generic 3D face model as a template that can be quickly adapted to different users through morphing, rather than creating detailed 3D models from scratch for each user. This copying approach maintains realism while significantly reducing model generation time

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent pre-generates shadow maps and lighting information on the generic 3D face model before the actual virtual try-on. These pre-computed shadow transforms are then applied to 2D face images, eliminating the need for time-consuming ray tracing during the actual try-on process

Inventive Principle:

Principle #10Preliminary action

2Manufacturing precision

If a detailed 3D model of the user's face with texture is generated, then the accuracy of shadow rendering is improved, but the complexity of the process and computational resources required increase

Engineering Contradiction:

Improveaccuracy of shadow renderingVSAvoidcomplexity of model generation process

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent separates the shadow generation process into independent components: a generic 3D face model for shadow map generation, and 2D face images for final composition. This segmentation allows shadow rendering to be performed on a simplified model while maintaining accuracy when applied to actual user images

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces a generic 3D face model as an intermediary between the lighting/glasses model and the user's actual face images. This intermediary model serves as a mediator that captures shadow information without requiring detailed user-specific geometry, simplifying the overall process

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS10403036B2Rendering glasses shadows

Publication Date: 2019.09.03 DITTO TECH INC

AI summary

Rendering glasses with shadows is disclosed, including: generating a face image corresponding to an image of a set of images based at least in part on a face model, wherein the set of images is associated with a user's face; generating a face with shadows image corresponding to the image based at least in part on shadows casted by a glasses model on the face model; generating a shadow transform based at least in part on a difference determined based at least in part on the face image and the face with shadows image; generating a shadowed image based at least in part on applying the shadow transform to the image; and presenting the shadowed image including by overlaying a glasses image associated with the glasses model over the shadowed image.