Gaussian Receptive Field Face Recognition for Mobile Devices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional face recognition systems are impractical for handheld and mobile devices due to high computational and memory requirements, as well as uncontrolled imaging conditions, which lead to inaccurate computations and high costs.
Innovation Solution
A facial recognition system utilizing hierarchical feature learning and large-scale classification engine training with Gaussian Receptive Fields, over-complete subset theory features, and a support vector machine for efficient image classification, enabling low-cost and accurate facial recognition on devices with limited resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional face recognition systems are implemented, then recognition accuracy can be achieved, but computational complexity and memory requirements become excessively high
Solution Approach 1:
The patent segments the face recognition process into distinct hierarchical stages: local feature extraction using Gaussian receptive fields, feature selection based on discriminative power, and classification using support vector machines. This segmentation allows each stage to be optimized independently, reducing overall computational complexity while maintaining accuracy
Solution Approach 2:
The patent extracts only the most discriminative local features from face images using Gaussian receptive fields, rather than processing entire images or all possible features. This selective extraction significantly reduces the data volume requiring computation and storage, addressing the memory and computational constraints of mobile devices
2Measurement precision
If conventional face recognition systems are implemented, then recognition functionality is provided, but memory footprint becomes too large for handheld devices
Solution Approach 1:
The patent extracts only the most discriminative local features from face images using Gaussian receptive fields, rather than processing entire images or all possible features. This selective extraction significantly reduces the data volume requiring computation and storage, addressing the memory and computational constraints of mobile devices
Solution Approach 2:
The patent employs partial action by selecting only the top k most discriminative features from the extracted feature set, rather than using all extracted features. This partial utilization of features reduces memory requirements while maintaining recognition accuracy, making the system feasible for resource-constrained handheld devices
3Measurement precision
If conventional face recognition systems are implemented, then recognition capability is achieved, but cost becomes prohibitive for mobile applications
Solution Approach 1:
The patent employs computationally efficient algorithms with low computational complexity requirements, making the system suitable for implementation on inexpensive mobile devices with limited processing power. The approach trades off some computational intensity for broader hardware compatibility, reducing implementation costs
Solution Approach 2:
The patent segments the face recognition process into distinct hierarchical stages: local feature extraction using Gaussian receptive fields, feature selection based on discriminative power, and classification using support vector machines. This segmentation allows each stage to be optimized independently, reducing overall computational complexity while maintaining accuracy
4Adaptability or versatility
If face recognition is performed under uncontrolled imaging conditions, then real-world applicability is improved, but computation accuracy deteriorates
Solution Approach 1:
The patent applies local quality by using Gaussian receptive fields with different parameters to extract features from different local regions of the face image. This allows the system to adapt to varying imaging conditions in different regions while maintaining overall recognition accuracy, as each local feature is extracted with appropriate sensitivity to its specific context
Data Source
AI summary
Methods and systems may provide for facial recognition of at least one input image utilizing hierarchical feature learning and pair-wise classification. Receptive field theory may be used on the input image to generate a pre-processed multi-channel image. Channels in the pre-processed image may be activated based on the amount of feature rich details within the channels. Similarly, local patches may be activated based on the discriminant features within the local patches. Features may be extracted from the local patches and the most discriminant features may be selected in order to perform feature matching on pair sets. The system may utilize patch feature pooling, pair-wise matching, and large-scale training in order to quickly and accurately perform facial recognition at a low cost for both system memory and computation.