Partial Image Transmission for IoT Face Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing cloud-based communication systems for IoT face delays due to increased network loads when continuously transmitting image data for face recognition, leading to delays in verification results.
Innovation Solution
A distributed processing system where an image acquisition terminal captures images, detects objects, and transmits partial images to a distributed data processing terminal, which then forwards the data to a centralized server for verification, reducing network loads and improving response times.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the image acquisition terminal continuously transmits image data to the server, then the face recognition processing can be performed, but the network load increases causing response delay
Solution Approach 1:
The patent segments the image data transmission by extracting and transmitting only the face portion (partial image) rather than the entire captured image. The object detector identifies face regions in the captured image, and only these extracted face portions are transmitted to the server for verification, reducing the data volume and network load while maintaining recognition functionality
Solution Approach 2:
The patent extracts the essential verification data (face portion) from the complete captured image. By using an object detector to identify and extract only the relevant face regions, the system removes unnecessary background and non-face elements, transmitting only the critical information needed for verification to reduce network bandwidth consumption
2Measurement precision
If the verification process extracts feature values from received image data, then accurate verification can be performed, but the server load increases causing delay in transmitting verification results
Solution Approach 1:
The patent applies segmentation by dividing the verification process into two stages: local preliminary verification (extracting feature values from transmitted partial images) and server-based final verification. This segmentation distributes the processing load, with the terminal handling initial feature extraction and the server performing definitive verification, reducing overall processing time and server burden
Data Source
AI summary
A communication system includes: an image acquisition terminal to obtain a captured image, the captured image including at least one object that is captured with an imaging unit; a distributed data processing terminal to display the captured image; and a centralized data processing server connected with the distributed data processing terminal through a network. The image acquisition terminal includes: an object detector to detect the object in the captured image to generate a partial image that is a part of the captured image having the detected object; and a first transmitter to transmit the partial image to the distributed data processing terminal as data to be verified. The distributed data processing terminal includes: a second transmitter to transmit the data to be verified that is received from the image acquisition terminal, and verification data that is registered, respectively, to the centralized data processing server, the verification data being an image representing a specific object. The centralized data processing server includes: a verification unit to verify the data to be verified, using the verification data that is received from the distributed data processing terminal, to generate a verification result; and a third transmitter to transmit the verification result to the distributed data processing terminal, to cause the distributed data processing terminal to display information based on the verification result with the captured image.