Face Recognition Search via Segmented Mobile Processing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current face recognition and social networking profile searching systems lack efficient methods for mobile devices to identify and match faces in images with social networking profiles, particularly in real-time scenarios, and do not effectively integrate facial recognition with social networking services to provide relevant results.

Innovation Solution

A computer-implemented system that uses a mobile device with a camera and screen to interact with a social networking service, where a search service compares face representations from an image with member profiles, providing results that match search terms and facial features, and displays matches on the device screen, optionally including friend identifiers and advertisement information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If face recognition is performed on mobile devices in real-time, then user interaction speed is improved, but device processing power and energy consumption increase

Engineering Contradiction:
Improvereal-time face recognition speedVSAvoidmobile device energy consumption
Core Design Contradiction:
SpeedVSUse of energy by moving object

Solution Approach 1:

The system divides the face recognition task into segments: the mobile device captures images and performs initial face detection, then transmits only the detected face data to remote servers for comprehensive matching against social networking profiles. This segmentation allows real-time response while minimizing local energy consumption.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary server system that acts as a mediator between the mobile device and social networking databases. The server handles the computationally intensive tasks of comparing face representations with stored profiles, reducing the energy burden on the mobile device while maintaining real-time interaction capability.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If comprehensive social networking profile searching is performed, then result accuracy is improved, but search time and system complexity increase

Engineering Contradiction:
Improveface matching accuracyVSAvoidsearching system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

Social networking profiles are pre-indexed with facial recognition data and relevant profile information before search queries are executed. This preliminary organization of data allows the system to perform accurate comprehensive searches without requiring complex real-time processing, thereby reducing system complexity while maintaining high accuracy.

Inventive Principle:
Principle #10Preliminary action

3Reliability

If multiple face representations are processed simultaneously, then recognition completeness is improved, but processing time and computational load increase

Engineering Contradiction:
Improveface recognition completenessVSAvoidprocessing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system dynamically adjusts processing priorities based on the number of faces detected. When multiple faces are present, the system processes them in prioritized batches, focusing first on faces that meet confidence thresholds, and progressively processing additional faces based on available resources and time constraints, thereby maintaining recognition completeness while managing processing time.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS8917913B2Searching with face recognition and social networking profiles
Publication Date: 2014.12.23 WRP IP MANAGEMENT LLC
  • US8917913B2 patent drawing
  • US8917913B2 patent drawing
  • US8917913B2 patent drawing

AI summary

Systems for performing on-line searching, and particularly, to searching with face recognition and social networking profiles. In one example, one or more systems may be provided with regard to searching with face recognition and social networking profiles.