Real-Time Video Analysis for Dynamic Individual Identification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for identifying individuals in a crowd rely heavily on memory and visual cues, which are inadequate for recognizing unfamiliar people or those with similar appearances, especially when characteristics like national origin, demographic, or health status are considered, limiting effective identification.

Innovation Solution

Implementing real-time video analysis on mobile devices using object recognition technology to match facial features, symmetry, eye color, hair color, body type, clothing, and other identifiable traits from a video stream with stored images, including social networking and transaction data, to dynamically identify individuals and provide additional information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If real-time video analysis with object recognition is implemented, then identification accuracy is improved, but device complexity increases

Engineering Contradiction:
Improveidentification accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent employs an intermediary processing system that captures video streams, extracts facial features, and matches them against databases. This intermediary layer handles the complex analysis tasks, allowing the mobile device to achieve high identification accuracy without requiring the entire system complexity to be embedded in the end-user device. The intermediary server or cloud-based system performs the heavy computational lifting.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The identification system is segmented into multiple independent components: video capture module, facial feature extraction module, database matching module, and result presentation module. Each component handles a specific aspect of the identification process, allowing for optimized processing and reducing the complexity burden on any single device. The segmentation enables distributed computing where complex tasks are divided across multiple processing units.

Inventive Principle:
Principle #1Segmentation

2Reliability

If multiple identification criteria are analyzed in real-time, then identification reliability is improved, but processing time increases

Engineering Contradiction:
Improveidentification reliabilityVSAvoidprocessing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary actions by pre-processing and storing facial feature data in databases before actual identification is needed. Facial features are extracted and stored in advance, allowing for rapid matching during real-time identification. This preliminary preparation enables the system to evaluate multiple criteria quickly during the actual identification process without excessive processing delays.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The video analysis operates continuously, with the system constantly processing video streams and maintaining ready-state identification capabilities. Multiple frames are analyzed in sequence, and the system continuously updates identification results as new information becomes available. This continuous processing allows the system to evaluate multiple identification criteria over time rather than requiring all analysis to complete in a single discrete step.

Inventive Principle:
Principle #20Continuity of useful action

Data Source

PatentUS8611601B2Dynamically indentifying individuals from a captured image
Publication Date: 2013.12.17 BANK OF AMERICA CORP
  • US8611601B2 patent drawing
  • US8611601B2 patent drawing
  • US8611601B2 patent drawing

AI summary

Embodiments of the invention are directed to methods and apparatuses for capturing a real-time video stream using a mobile device, determining, using a processor, which images from the real-time video stream are associated with individuals meeting a user defined criteria, and presenting on a display of the real-time video stream, one or more indicators, each indicator being associated with an image determined to be a person meeting the predefined criteria.