Driver Monitoring System rPPG Adaptive Interpolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driver monitoring systems (DMS) face challenges in real-time performance due to insufficient computing power and inaccuracies in remote-PhotoPlethysmoGraphy (rPPG) processes, particularly caused by bad data from failed detection or shading on the driver's face.
Innovation Solution
The proposed method involves obtaining sequences of frames from face images, determining the presence of bad color scalars, and performing adaptive interpolation based on their locations. If bad color scalars are at the beginning or end of an array, the entire array is discarded to prevent erroneous estimations and conserve computing power.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing rPPG process performs interpolation for all bad data, then measurement precision is improved, but computing power consumption increases and real-time performance is compromised
Solution Approach 1:
The patent applies different handling strategies to different locations of bad data based on their positions in the array. Bad data at the beginning or end of the array is discarded, while bad data in the middle undergoes interpolation. This localized differentiation optimizes both computing efficiency and measurement accuracy by avoiding unnecessary interpolation operations.
Solution Approach 2:
Instead of performing interpolation for all bad data points (excessive action), the patent applies interpolation only to bad data points that are not at the beginning or end of the array (partial action). This selective approach reduces computational load while maintaining sufficient measurement precision.
2Measurement precision
If existing rPPG process performs interpolation to make up for bad data, then measurement precision is improved, but loss of time increases due to additional processing steps
Solution Approach 1:
The patent differentiates the handling of bad data based on their locations. By discarding bad data at the beginning or end of the array and only interpolating bad data in the middle, the processing time is significantly reduced while maintaining measurement precision for the most critical data points.
Solution Approach 2:
The patent applies partial interpolation only where necessary (in the middle of the array) rather than performing full interpolation across the entire array, thereby reducing processing time while maintaining sufficient accuracy.
3Measurement precision
If existing rPPG process uses conventional interpolation algorithms, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent simplifies the overall algorithm complexity by implementing a location-based decision framework that discards bad data at the beginning or end and only applies interpolation to middle sections. This structured approach reduces the complexity compared to applying universal interpolation algorithms to all data points.
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
This approach enhances the computing speed and accuracy of the rPPG algorithm, enabling real-time performance in DMS by reducing the computational load and excluding erroneous data from the interpolation process.
Implementation Method 1
The rPPG is measuring the contrast between specular reflection and diffused reflection. The specular reflection is the pure light reflection from the skin
Implementation Method 2
the diffused reflection is the reflection that remains from the absorption and scattering in the skin tissue, which varies as blood volume changes
Implementation Method 3
the existing rPPG process usually first determines whether the collected data contains bad data. If it determines that there is bad data, the existing rPPG process usually performs interpolation to make up for all the bad ones
Data Source
AI summary
The disclosure describes a method adapted for a driver monitoring system and the driver monitoring system. The method may comprise obtaining sequences of frames, wherein each sequence of frames consists of N arrays of color scalars corresponding to N face zones of a driver. For each array, the method may determine whether bad color scalars exist in the array; and perform interpolation adaptively to the array based on locations of bad color scalars in the array, in response to the determination of bad color scalars existing in the array.