Age Progression Video Generation Using Sliding Window Facial Recognition

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing systems for creating age progression videos from collections of photos are limited in their ability to automatically identify and sequence images of a person across a wide age range, often requiring manual selection and lacking integration with remote storage and facial recognition algorithms for efficient video generation and sharing.

Innovation Solution

A media processing engine that uses a sliding window for facial recognition across a wide age range, implemented on a remote server, to identify and sequence photos of a person from a chronological perspective, allowing for automatic creation and sharing of age progression videos, which can include viral marketing elements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a sliding window approach is used for facial recognition across a wide age range, then the accuracy of identifying the same person across different ages is improved, but the complexity of the recognition algorithm increases

Engineering Contradiction:
Improvefacial recognition accuracyVSAvoidalgorithm complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent divides the facial recognition task into multiple time periods using a sliding window approach. Instead of attempting to recognize the same person across their entire lifespan in one step, the system segments the recognition process into overlapping time windows (e.g., 0-10 years, 5-15 years, 10-20 years), making each recognition task more manageable and accurate while reducing overall algorithmic complexity.

Inventive Principle:
Principle #1Segmentation

2Ease of operation

If photos are automatically selected and sequenced by the processing engine, then the ease of operation is improved, but the risk of selecting inappropriate photos increases

Engineering Contradiction:
Improveautomatic video creationVSAvoidphoto selection accuracy
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The system implements feedback mechanisms where the processing engine continuously evaluates selected photos against multiple criteria (quality scores, chronological consistency, facial recognition confidence levels). Photos that don't meet the thresholds are rejected and replaced, ensuring high reliability in the final video while maintaining ease of operation for the user.

Inventive Principle:
Principle #23Feedback

3Quantity of substance

If the system processes a large number of photos to create a comprehensive age progression video, then the completeness of the video is improved, but the processing time increases

Engineering Contradiction:
Improvenumber of photos processedVSAvoidvideo generation time
Core Design Contradiction:
Quantity of substanceVSLoss of time

Solution Approach 1:

The patent applies partial action by processing only the necessary subset of photos rather than all available photos. The sliding window approach and quality-based filtering ensure that only photos within relevant time periods and meeting quality thresholds are processed, reducing overall processing time while maintaining video completeness.

Inventive Principle:
Principle #16Partial or excessive action

4Adaptability or versatility

If the sliding window time period is made larger to cover more age range, then the coverage of age progression is improved, but the precision of chronological association decreases

Engineering Contradiction:
Improveage range coverageVSAvoidchronological association accuracy
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The system segments the age progression coverage into multiple overlapping time windows. Each window maintains high chronological precision for its specific range, while the overlapping nature of consecutive windows ensures comprehensive age range coverage. For example, windows of 0-10 years, 5-15 years, and 10-20 years overlap to both cover the full 0-20 year range and maintain precision within each segment.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3007173B1Apparatus, system and method for automatically creating an age progression video
Publication Date: 2020.04.22 SYNCHRONOSS TECHNOLOGIES INC
  • EP3007173B1 patent drawingFigure 1
  • EP3007173B1 patent drawingFigure 2
  • EP3007173B1 patent drawingFigure 3

AI summary

Embodiments of the present invention relate to a media processing engine that produces an age progression video of a person by using photos of the person. The processing engine sequences and manipulates the photos such that the focus in each chronological photo is on the person. The processing engine uses a sliding window for facial recognition across a wide age range such that pictures of the person at a younger age (e.g., infancy) are positively associated with pictures of the person at an older age (e.g., teenager). In some embodiments, the processing engine is implemented on a remote server that provides a backup or storage service to its members. The photos are located at the remote server. The age progression video can be downloaded and/or shared by a member with others. In some embodiment, the age progression video includes a text caption, such as a message promoting the service.