Multi-Algorithm Face Recognition with Parallel Dataset Partitioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing face recognition systems face challenges in accurately recognizing faces due to variations in pose, illumination, and other environmental factors, and often rely on single algorithms that are not optimal for all conditions, leading to reduced recognition accuracy.
Innovation Solution
A multi-phase face recognition system that normalizes images, generates distance matrices, and uses a MapReduce program for parallel processing across multiple algorithms, including Eigenface, LDA, ICA, and Gabor-Wavelet transforms, to enhance recognition accuracy by integrating multiple algorithms in a cloud-based environment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single face recognition algorithm is used, then the system complexity is low, but the recognition accuracy deteriorates under variations in pose and illumination
Solution Approach 1:
The patent combines multiple face recognition algorithms (Eigenface, Fisherface, and other classification algorithms) into a single integrated system. Each algorithm processes the input image independently, and their results are merged through voting or weighted combination to produce the final recognition outcome, thereby improving accuracy while managing complexity through modular architecture
Solution Approach 2:
The system implements a universal face recognition framework that can adapt to different conditions (variations in pose, illumination, and environment) by selecting and applying multiple algorithms appropriate for different scenarios. This multi-functional approach allows the system to maintain high recognition accuracy across diverse operating conditions
2Reliability
If multiple algorithms are used in parallel, then the recognition accuracy improves, but the processing time increases
Solution Approach 1:
The patent divides the face recognition process into distinct segments or stages, where different algorithms are applied to different subsets of the image data or different feature representations. This segmentation allows parallel processing of multiple algorithms simultaneously, reducing overall processing time while maintaining the accuracy benefits of multiple algorithms
Solution Approach 2:
The system implements a staged processing approach where not all algorithms are applied to all images. Instead, based on initial processing results or image characteristics, only the necessary subset of algorithms is applied, avoiding unnecessary computational overhead while maintaining high recognition accuracy for critical cases
3Reliability
If the entire dataset is processed by each algorithm, then the recognition accuracy is maximized, but the computational complexity and processing time increase significantly
Solution Approach 1:
The patent segments the large face image dataset into multiple subsets or partitions. Different algorithms process different subsets of the data in parallel, or the data is divided into training and testing portions that are processed efficiently. This segmentation reduces the computational burden on each algorithm while maintaining overall recognition accuracy through comprehensive data coverage
Solution Approach 2:
The system performs preliminary processing of the dataset, such as pre-computing feature extractions, creating index structures, or pre-processing images into standardized formats before the actual recognition process. This preliminary action reduces the computational complexity during runtime processing, allowing multiple algorithms to operate more efficiently on the pre-prepared data
Data Source
AI summary
A system and method of face recognition comprising multiple phases implemented in a parallel architecture. The first phase is a normalization phase whereby a captured image is normalized to the same size, orientation, and illumination of stored images in a preexisting database. The second phase is a feature extraction/distance matrix phase where a distance matrix is generated for the captured image. In a coarse recognition phase, the generated distance matrix is compared with distance matrices in the database using Euclidean distance matches to create candidate lists, and in a detailed recognition phase, multiple face recognition algorithms are applied to the candidate lists to produce a final result. The distance matrices in the normalized database may be broken into parallel lists for parallelization in the feature extraction/distance matrix phase, and the candidate lists may also be grouped according to a dissimilarity algorithm for parallel processing in the detailed recognition phase.