Network Camera Local Similarity Matching for Accurate Face Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The integration of network cameras with back-end servers for face comparison often faces hardware and software discrepancies, leading to unsuccessful comparisons or low accuracy in face image analysis.
Innovation Solution
The network camera processes images using its own processor, storing them in a targeted manner based on similarity thresholds and assigning unique device target identifiers, reducing the need for server-side comparisons and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If face images are acquired by network cameras and sent to back-end servers for comparison, then the system can perform centralized face recognition, but hardware and software differences between different network cameras and servers cause comparison failures or low accuracy
Solution Approach 1:
The patent applies preliminary action by having the network camera perform face image processing and extraction before sending data to the server. The camera extracts key facial features and pre-processes images according to a unified protocol, ensuring that the server receives standardized data regardless of hardware differences. This pre-processing step resolves compatibility issues by standardizing the input data format before transmission.
Solution Approach 2:
The patent introduces an intermediary protocol that standardizes the interface between network cameras and back-end servers. This protocol acts as a mediator that translates different camera hardware outputs into a unified format, enabling seamless communication and comparison across diverse platforms without requiring the server to adapt to each specific camera type.
2Ease of operation
If all face images are stored and compared on the back-end server, then centralized management is achieved, but the system complexity and processing load on the server increase significantly
Solution Approach 1:
The patent applies segmentation by dividing the face recognition system into two functional parts: the network camera handles local image acquisition, preprocessing, and feature extraction, while the back-end server handles only the comparison and database management. This segmentation reduces the server's processing burden and distributes computational tasks appropriately across the system architecture.
Solution Approach 2:
The patent extracts the computationally intensive image processing and feature extraction functions from the back-end server and relocates them to the network camera. This extraction reduces the server's workload to only essential comparison operations, thereby simplifying server complexity while maintaining centralized management capabilities through standardized data transmission.
3Reliability
If face comparison is performed on the back-end server, then centralized processing is implemented, but the accuracy of comparison results decreases due to hardware and software differences
Solution Approach 1:
The patent applies preliminary action by having the network camera extract standardized facial features and pre-process images before transmission. This ensures that the comparison data sent to the server is already optimized and standardized, improving comparison accuracy while avoiding the need for the server to handle hardware-specific variations.
Solution Approach 2:
The patent changes the parameters of image transmission by converting raw images into standardized feature vectors with unified formatting and protocols. This parameter transformation ensures that all data sent to the server conforms to consistent specifications, thereby improving comparison accuracy regardless of the original camera hardware differences.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
The application provides a network camera, a video monitoring system and a method. The network camera comprises: an image sensor, a processor, a memory and a network communication interface; the processor is configured for matching a current image acquired by the image sensor with an image stored in a second storage unit of the memory, so as to obtain a similarity value representing a matching result, and storing an image satisfying a similarity condition in another storage unit of the network camera, reducing the difficulty of image comparison and improving the accuracy of the comparison result.