Microlens Array Image Processing for Fingerprint Authentication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Microlens array imaging devices face high processing loads and inefficiencies in calculating movement vectors for fingerprint authentication due to the need to generate whole images from multiple inverted and overlapping microlens photographs, leading to increased processing time and load.
Innovation Solution
An image processing device and method that calculates movement vectors directly from consecutive microlens photographs without generating whole images, using a movement vector calculating section to determine vectors based on feature points within single, two-image duplicate, and four-image duplicate regions, and an image synthesizing section to align and combine images for high-quality fingerprint authentication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If whole images are generated by combining multiple microlens photographing images before movement vector calculation, then alignment and synthesis processing can be performed, but processing time and processing load become enormous
Solution Approach 1:
The patent divides the image processing into segments by performing movement vector calculation on individual microlens photographing images before combining them. Instead of generating a complete whole image first and then calculating movement vectors, the system calculates movement vectors for each microlens image segment separately, significantly reducing the computational burden and processing time while maintaining alignment precision.
Solution Approach 2:
The patent performs preliminary movement vector calculation on microlens photographing images before the final image synthesis. By calculating movement vectors in advance on the segmented microlens images rather than on the final combined image, the system prepares alignment data earlier in the processing pipeline, reducing the overall processing time and enabling more efficient image synthesis.
2Manufacturing precision
If vertical and horizontal inversion processing and image cutout are performed for each microlens photographing image, then accurate alignment can be achieved, but processing load becomes enormous
Solution Approach 1:
The patent extracts only the necessary processing steps (movement vector calculation) from the complete image processing pipeline. Instead of performing full inversion processing and image cutout for each microlens photographing image, the system extracts and processes only the relevant movement vector information, reducing processing complexity while maintaining alignment precision.
3Measurement precision
If multiple pieces of processing are performed for each frame unit to generate high quality synthetic images, then image quality improves, but processing load becomes enormous
Solution Approach 1:
The patent segments the image processing workflow to perform movement vector calculation on individual microlens photographing images before synthesis. This segmentation allows the system to maintain high image quality through precise alignment while avoiding the need to perform multiple complex processing operations on complete whole images, thereby improving processing efficiency.
Solution Approach 2:
The patent performs partial processing by calculating movement vectors only on the necessary microlens photographing images rather than performing complete processing on all images. This partial action approach maintains sufficient alignment precision for high-quality synthesis while reducing the overall processing load and improving efficiency.
Data Source
AI summary
A device and a method are provided which perform movement vector calculation processing from photographing images for which a microlens array is used without generating whole images. Specifically, consecutive photographing images photographed by an imaging section having a microlens array are input, and a movement vector between the images is calculated. A movement vector calculating section calculates the movement vector corresponding to a feature point by using microlens photographing images photographed by respective microlenses. When the movement vector corresponding to the feature point within duplicate image regions having the same image region within a plurality of microlens photographing images is calculated, an average value of a plurality of movement vectors corresponding to the same feature point is calculated and is set as the movement vector of the feature point. Alternatively, an average value of plural movement vectors remaining after an outlier (abnormal value) is excluded is calculated and is set as the movement vector of the feature point.