Flash Control Using Statistical Face Models to Prevent Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Digital cameras lack effective flash control mechanisms to prevent black or white compression and contrast issues caused by shadows, especially since they do not have special hardware like silver-salt film cameras, which can correct such issues during printing.
Innovation Solution
The implementation of a photography apparatus and method that uses imaging means, parameter acquisition means, and auxiliary light source control to adjust lighting conditions based on statistical characteristic quantities obtained through models fitted to images of a subject, allowing for appropriate flash control without special hardware.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If special hardware like shadow detection devices under glass windows is used to predict shadow appearance and control flash, then flash control accuracy is improved, but device complexity increases
Solution Approach 1:
The patent uses a pre-built statistical model (copy of shadow behavior patterns) obtained through statistical processing of multiple images, replacing the need for special shadow detection hardware. The model captures shadow characteristics under various lighting conditions and is applied to control the flash based on ordinary image data, achieving accurate flash control without additional measurement devices.
Solution Approach 2:
The patent replaces the mechanical/optical shadow detection system (glass window with black mark and imaging device) with a computational approach using statistical models and image processing. The flash control is achieved through software-based analysis of ordinary images rather than specialized hardware measurement, substituting physical measurement mechanisms with information processing.
2Loss of information
If digital camera photography is performed without adequate flash control, then information loss occurs due to black or white compression, but adding special hardware increases device complexity
Solution Approach 1:
The patent uses a statistical model copied from training data (multiple images under different lighting conditions) to predict shadow behavior and optimize flash control. This model-based approach prevents information loss by anticipating lighting issues before they occur, without requiring special sensors or hardware beyond the ordinary imaging device.
Solution Approach 2:
The patent performs preliminary statistical processing on multiple images to build a shadow prediction model before actual photography. This pre-computed model enables the camera to predict shadow appearance and adjust flash settings in advance, preventing black or white compression before it occurs rather than correcting it after image capture.
3Measurement precision
If statistical processing on multiple images is performed to build lighting condition models, then flash control accuracy is improved, but processing time increases
Solution Approach 1:
The patent performs the computationally intensive statistical processing and model building in advance, during a preliminary training phase using multiple reference images. Once the statistical model is constructed, it can be quickly applied to control flash in subsequent photography operations, separating the time-consuming model creation from the real-time image capture process.
Solution Approach 2:
The patent uses statistical processing on a representative subset of images (multiple images under different lighting conditions) to build the model, rather than processing all possible images. This partial action approach achieves sufficient prediction accuracy without requiring exhaustive processing of every possible lighting scenario, balancing model quality with processing time.
Data Source
AI summary
An auxiliary light source such as a flash is controlled for highly accurately reflecting a lighting condition with less effect of a factor other than the lighting condition such as contrasting density caused by shadow or black compression or white compression. For this purpose, a parameter acquisition unit obtains weighting parameters for principal components representing lighting conditions in a face region in an image detected by a face detection unit, by fitting to the detected face region a mathematical model generated according to a method of AAM using a plurality of sample images representing human faces in different lighting conditions. A flash control unit obtains a flash control condition by referring to a reference table according to the parameters, and controls flash light emission according to the control condition.