Facial ROI Expansion for Accurate Luma-Based Exposure
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image capture devices face issues with unrealistic skin tones and unappealing skin texture due to overhead lighting, as traditional facial region of interest calculations exclude the forehead region, leading to inaccurate exposure parameters.
Innovation Solution
Expand the facial region of interest to include the forehead region by identifying key facial landmarks and using a segmentation mask to distinguish skin areas, adjusting the initial region of interest to encompass these points, and calculating luma based on skin pixels within the expanded area.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a small region of interest (from mouth to eyes) is used to calculate facial luma, then the calculation is simple and fast, but the exposure parameters create unrealistic skin tones and unappealing skin texture under overhead lighting
Solution Approach 1:
The patent segments the facial region of interest into multiple zones: the traditional mouth-to-eyes region and the additional forehead region. By dividing the face into these segments and calculating luma for each, the system achieves more accurate overall exposure determination while maintaining manageable computational complexity through focused processing of each segment.
Solution Approach 2:
The patent extends the region of interest upward into the forehead dimension, adding a new spatial dimension to the traditional horizontal mouth-to-eyes measurement. This dimensional expansion captures overhead lighting effects that were previously missed, improving luma calculation accuracy without requiring completely new processing methodologies.
2Reliability
If the region of interest is expanded to include the forehead, then exposure accuracy under overhead lighting improves, but the computational complexity increases
Solution Approach 1:
The patent applies different processing qualities to different facial regions. The forehead region receives specialized handling with axis-based point identification and selective inclusion, while the traditional mouth-to-eyes region maintains standard processing. This localized quality approach ensures reliable exposure under overhead lighting without uniformly increasing complexity across the entire facial processing pipeline.
3Measurement precision
If skin segments are identified using segmentation mask within the expanded region, then the luma calculation becomes more accurate, but the processing time increases
Solution Approach 1:
The patent performs preliminary identification of skin segments using segmentation masks before the actual luma calculation. By pre-segmenting the facial regions and identifying skin areas in advance, the system prepares the data structure for efficient luma computation, reducing the time penalty of the expanded processing region through upfront organization of the facial data.
Data Source
AI summary
A method of determining a facial luma includes generating, based on image data, facial landmark data and facial region of interest data using the one or more machine learning models. The facial landmark data is indicative of facial landmarks of a face in the scene, and the facial region of interest data is indicative of an initial region of interest of the face. The method includes generating an adjusted region of interest of the face by expanding the initial region of interest. The method also includes determining an exposure of an image of the scene based on the adjusted region of interest.