Automatic Image Duplicate Detection Using Segmented Checksum and Fingerprint Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for detecting duplicate images in large collections are inefficient as they require user intervention and fail to account for subjective user perception of image similarity, leading to unnecessary computation time and incorrect duplicate identification.

Innovation Solution

A method that classifies transformed images based on user perception by determining fingerprint distance thresholds and associating actions with each classification zone, allowing for automatic management of images by distinguishing between identical, slightly modified, and largely modified copies.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If fingerprint technique is used to detect near-duplicate images with various transformations, then the range of duplicate detection is improved, but the computation time increases significantly (up to 500 times slower than checksum)

Engineering Contradiction:
Improverange of duplicate detectionVSAvoidcomputation time
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent segments the duplicate detection process into two distinct stages: first using checksum for rapid filtering of exact duplicates, then using fingerprint technique only on the remaining non-duplicate images to detect near-duplicates. This segmentation allows the system to benefit from both methods without paying the full computational cost of fingerprinting all images.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies the computationally expensive fingerprint technique partially - only to images that were not already identified as duplicates by the faster checksum method. This partial application of the slower method reduces overall computation time while still achieving comprehensive duplicate detection.

Inventive Principle:
Principle #16Partial or excessive action

2Loss of time

If checksum technique is used to detect exact duplicates, then the computation time is reduced, but the ability to detect near-duplicate images with transformations is lost

Engineering Contradiction:
Improvecomputation timeVSAvoidrange of duplicate detection
Core Design Contradiction:
Loss of timeVSAdaptability or versatility

Solution Approach 1:

The patent divides the detection task into two segments: checksum handles exact duplicates (bit-by-bit identical), while fingerprint handles near-duplicates (visually similar but with transformations). Each method is applied to its appropriate segment of the problem space.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The checksum method acts as an intermediary filtering step that prepares the data for the more sophisticated fingerprint analysis. By first removing exact duplicates through checksum, the system reduces the workload for the fingerprint technique and ensures it focuses only on potentially interesting cases.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Reliability

If the fingerprint technique is tuned to detect severe transformations, then the robustness of duplicate detection is improved, but the computation time increases and false positives may occur for images that users do not consider duplicates

Engineering Contradiction:
Improverobustness of duplicate detectionVSAvoidcomputation time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent introduces dynamic, user-configurable thresholds that allow the fingerprint matching criteria to be adjusted based on user needs. Users can set minimum and maximum fingerprint distances to define their personal definition of duplicate, making the system adaptive rather than fixed.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the parameters of the fingerprint technique by allowing users to configure distance thresholds. This parameter adjustment enables the same fingerprint algorithm to serve different user needs - from strict duplicate detection to more permissive near-duplicate detection - without changing the core algorithm.

Inventive Principle:
Principle #35Parameter changes

4Measurement precision

If user intervention is required for each duplicate detection, then the accuracy of duplicate management is improved, but the automation level decreases and user time is consumed

Engineering Contradiction:
Improveaccuracy of duplicate managementVSAvoidautomation level
Core Design Contradiction:
Measurement precisionVSExtent of automation

Solution Approach 1:

The patent enables the system to serve itself by automatically applying user-defined rules to manage duplicates. Once users configure their preferences and thresholds, the system autonomously identifies and manages duplicates according to those preferences, reducing the need for continuous user intervention.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent incorporates feedback mechanisms where user responses to initial duplicate detections are used to refine and adjust the detection parameters. The system learns from user feedback and adapts its behavior, improving automation accuracy over time while reducing the frequency of user interventions needed.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP2608078B1Method of automatic management of images in a collection of images and corresponding device
Publication Date: 2020.03.04 INTERDIGITAL CE PATENT HOLDINGS SAS
  • EP2608078B1 patent drawingFigure 1
  • EP2608078B1 patent drawingFigure 2
  • EP2608078B1 patent drawingFigure 3

AI summary

The present invention relates to the field of management of image data in data storage. In particular, the present invention relates to a method and device for automatic detection of duplicate images in data storage and corresponding device while taking into account user perception of what he considers to be duplicate images, which method and device are particularly efficient with regard to the personalized, automatic management of large amounts of image data.