Monocular Face Scaling Using Defocus Depth Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing monocular cameras struggle to accurately determine the true scale of a user's face for applications like virtual try-on and augmented reality shopping due to image scale ambiguity and lack of depth information.
Innovation Solution
A method using a monocular camera to estimate face size by leveraging image defocus, employing autofocus or focal stacks, and applying a face mesh to improve accuracy, without requiring a dense depth map.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a monocular camera is used for face measurement, then device complexity is reduced, but measurement precision deteriorates due to image scale ambiguity and lack of depth information
Solution Approach 1:
The patent introduces an intermediary computational process that analyzes defocus blur in monocular images to infer depth information. By using the defocus amount as an intermediate indicator, the system can estimate face scale without requiring complex multi-camera setups or depth sensors, thus maintaining device simplicity while improving measurement precision.
Solution Approach 2:
The patent transforms the problem by changing the parameter being analyzed from sharp focus features to defocus blur characteristics. By measuring the amount of blur in different regions of the face image, the system can infer relative depth information and calculate face scale, effectively converting a limitation (defocus) into a useful measurement parameter.
2Measurement precision
If defocus information is used for depth estimation, then measurement precision improves, but device complexity increases due to autofocus or focal stacks requirements
Solution Approach 1:
The patent employs the camera's existing autofocus mechanism to serve the dual purpose of both capturing the image and generating defocus information for depth estimation. The autofocus system naturally creates varying degrees of defocus based on focus distance, and the patent leverages this self-generated defocus pattern without requiring additional hardware or complex focal stack mechanisms.
Solution Approach 2:
The patent recovers useful depth information from the previously discarded defocus blur. Instead of treating defocus as a degradation to be eliminated through precise focusing, the system deliberately utilizes the defocus blur as a carrier of depth information, transforming a waste product into a valuable measurement signal.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables precise face scaling for virtual try-on and augmented reality applications by accurately determining the user's face size using a single camera, enhancing the user experience.
Implementation Method 1
A method using a monocular camera to estimate face size by leveraging image defocus
Data Source
AI summary
A device that measures a size of a user's face, referred to as face scaling, using a monocular camera. Depth is calculated from sparse feature points. A face mesh is used to improve the estimation accuracy. A processing pipeline detects face features by applying a face landmark detection algorithm to find the important face feature points such as the eyes, nose, and mouth. The processing pipeline estimates feature points depth using depth obtained through image defocus. The processing pipeline further scales the face using an estimated depth of the face features.