Facial Geometry Capture Using Sequential Polarized Lighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current techniques for capturing high-resolution facial geometry and reflectance are either not instantaneous, costly, or fail to provide lighting-independent texture maps and specular reflectance information, often requiring complex setups and longer capture times.
Innovation Solution
A system using multiple cameras and lights arranged to capture evenly distributed specular reflections, with a controller managing sequential image capture and polarized lighting conditions, allowing for the generation of high-resolution geometry and reflectance data from a single image per camera, leveraging diffuse and specular photometric observations for energy minimization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple photographs from spherical lighting setups are used to capture reflectance properties, then lighting-independent texture maps and specular reflectance information are obtained, but capture time increases and equipment complexity increases
Solution Approach 1:
The system segments the capture process by using multiple camera subgroups that sequentially capture images under different lighting conditions. Each subgroup captures images with specific lights illuminated, allowing the system to gather comprehensive reflectance data through controlled segmentation of the capture sequence rather than requiring all lights to be on simultaneously for extended periods
Solution Approach 2:
The system implements periodic action by cycling through different lighting conditions in a structured sequence. The controller illuminates different subsets of lights in periodic cycles, with each cycle capturing images under specific lighting configurations. This periodic illumination pattern enables comprehensive reflectance measurement while maintaining rapid capture speed by efficiently utilizing time-multiplexed lighting sequences
2Measurement precision
If multiple photographs from spherical lighting setups are used to capture reflectance properties, then lighting-independent texture maps and specular reflectance information are obtained, but equipment complexity increases
Solution Approach 1:
The system applies universality by using a single spherical lighting structure to serve multiple functions: it provides both geometric shape information and reflectance properties through sequential imaging under different lighting conditions. The same physical setup that captures standard multi-view images is also used for reflectance measurement by controlling which lights are illuminated, eliminating the need for separate specialized equipment
Solution Approach 2:
The system changes parameters by controlling the illumination state of different light subsets rather than changing the physical structure. By dynamically adjusting which lights are on or off in the spherical array, the system creates different lighting conditions for reflectance measurement without adding physical complexity. This parameter-based control (light on/off states) simplifies the system compared to requiring multiple physical lighting structures
3Productivity
If sequential image capture by camera subgroups is used, then high-resolution geometry and reflectance data are captured near-instantaneously, but lighting control complexity increases
Solution Approach 1:
The lighting control system is segmented into multiple independent light subsets that can be controlled separately. Each subgroup of cameras corresponds to specific light combinations, allowing the controller to illuminate only the relevant lights for each capture sequence. This segmentation reduces the control complexity compared to managing all lights simultaneously while maintaining the ability to create diverse lighting conditions for rapid sequential capture
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 near-instant, cost-effective capture of high-resolution facial geometry and reflectance with precise surface detail, suitable for authoring high-quality digital human characters, using commodity hardware and minimizing capture time to less than a blink reflex.
Implementation Method 1
A first of the lights may be near a first of the cameras and a second of the lights may be near a second of the cameras. The controller may cause the first of the lights to light when it causes the second of the cameras to capture the image of the subject and may cause the second of the lights to light when the first of the cameras captures the image of the portion of the human subject.
Data Source
AI summary
A system may almost instantly capturing high-resolution geometry and reflectance data of a portion of a human subject. The system may include multiple cameras, each oriented to controllably capture an image of the portion of the human subject from a different location in space; multiple lights, each oriented to controllably illuminate the portion of the human subject from a location in space significantly different than the location in space of the other lights; and a controller. The controller may divide the cameras into subgroups with each subgroup of cameras containing at least one camera and with each camera belonging to only one of the subgroups; cause each subgroup of cameras to sequentially capture a single image of the portion of the human subject; and cause at least one of the lights to light while each subgroup of cameras captures a single image of the portion of the human subject. The system may include an image processing system that generates the high resolution geometry and reflectance data based on only one image from each camera.A polarizing optical element may be between each camera and the portion of the human subject. A polarizer filter may be between each light and the portion of the human subject. A controller may cause all of the cameras to simultaneously capture a single image of the portion of the human subject while the portion of the human subject is illuminated by all of the lights. The specular reflections from the portion of the human subject that are captured by one of the cameras may have a color distribution across the portion of the human subject that is different than the specular reflections from the portion of the human subject that are captured by another of the cameras.