Parallel CPU GPU Facial Recognition for Mobile Terminal Speed

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Mobile terminals face security risks due to slow facial recognition speeds, which can compromise personal information and property security when unauthorized users access the device.

Innovation Solution

The method involves utilizing both the central processing unit (CPU) and graphics processing unit (GPU) to perform feature matching and liveness detection in parallel, reducing the overall execution time of facial recognition by dividing the process into feature matching and liveness detection, and further optimizing by performing depth of field and blink detection in parallel.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If facial recognition is performed using traditional sequential processing, then authentication accuracy is maintained, but recognition speed is slow

Engineering Contradiction:
Improvefacial recognition speedVSAvoidexecution time
Core Design Contradiction:
SpeedVSLoss of time

Solution Approach 1:

The patent divides facial recognition into two independent modules: feature matching (performed by CPU) and liveness detection (performed by GPU). These modules are executed in parallel rather than sequentially, allowing simultaneous processing of face identification and authenticity verification, thereby reducing total execution time while maintaining accuracy

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a dual-processor architecture where CPU and GPU act as intermediaries for different processing tasks. The CPU handles feature extraction and matching while the GPU performs liveness detection through blink and depth analysis. This intermediary division allows optimal utilization of each processor's strengths and enables parallel execution

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If comprehensive liveness detection is performed, then anti-spoofing accuracy is improved, but processing time increases

Engineering Contradiction:
Improveliveness detection accuracyVSAvoiddetection time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

Liveness detection is segmented into specific sub-tasks (blink detection, depth of field detection) that are independently processed by the GPU. This segmentation allows comprehensive anti-spoofing verification without requiring sequential execution of all checks, as they can run in parallel

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent replaces traditional sequential mechanical processing with GPU-accelerated parallel processing. The GPU's ability to perform multiple calculations simultaneously substitutes for sequential CPU-based liveness checks, maintaining comprehensive detection accuracy while reducing processing time

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP3454255B1Facial recognition method and related product
Publication Date: 2021.10.27 GUANGDONG OPPO MOBILE TELECOMMUNICATIONS CORP LTD
  • EP3454255B1 patent drawingFigure 1~2
  • EP3454255B1 patent drawingFigure 3~4
  • EP3454255B1 patent drawingFigure 5~6

AI summary

Embodiments of the present disclosure relate to mobile terminal technologies, a facial recognition method and related products are provided. The method includes: acquiring a face image by a mobile terminal (100); extracting face feature information from the face image and matching the face feature information with a predetermined face feature template by a central processing unit (110) of the mobile terminal (100); performing liveness detection according to the face image by a graphics processing unit (120) of the mobile terminal (100) when the CPU (110) extracts the face feature information from the face image and matches the face feature information with the predetermined face feature template.