Drop Shadow Detection in Facial Recognition Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for processing shadows in facial recognition images are not entirely satisfactory as they process all images, including those without shadows, leading to inefficiencies.
Innovation Solution
A method that detects shadows by capturing and resizing images, normalizing pixel values, calculating variance, reducing image resolution, and constructing a feature vector to classify images as having or not having cast shadows, thereby determining if shadow processing is necessary.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If shadow processing is applied to all images, then shadow removal completeness is improved, but processing efficiency deteriorates
Solution Approach 1:
The patent applies preliminary action by performing shadow detection before shadow processing. The system first analyzes the captured image to determine whether shadows are present, and only then proceeds to apply shadow processing if needed. This preliminary detection step prevents unnecessary processing of images without shadows, thereby improving processing efficiency while maintaining shadow removal completeness for images that actually contain shadows.
2Reliability
If shadow processing is applied to all images, then shadow removal reliability is improved, but processing time increases
Solution Approach 1:
The patent implements preliminary shadow detection to identify images containing shadows before applying shadow processing. This preliminary action reduces processing time by avoiding unnecessary shadow processing on images without shadows, while ensuring that images with shadows receive the necessary processing to maintain removal reliability.
3Reliability
If shadow processing is applied to all images, then processing thoroughness is improved, but computational waste increases
Solution Approach 1:
The patent applies preliminary shadow detection to identify images requiring shadow processing. This preliminary step reduces computational waste by directing processing resources only to images that actually contain shadows, while maintaining thorough processing for those images. The detection mechanism analyzes image characteristics to determine shadow presence, enabling selective processing that eliminates unnecessary computational expenditure.
Data Source
Figure 1~4
Figure 3
AI summary
The invention concerns a method (300) for detecting drop shadows on an image which consists of successively reducing the resolution of the captured image, calculating the variance of same and deducing, from said variances, whether or not drop shadows are present.