Image Annotation System Using Tag Co-occurrence Probability

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The existing image annotation processes are largely manual, especially for multi-tag annotations, which require significant manpower and resources, especially for images with many features.

Innovation Solution

An annotation method and device that determines a first probability value for annotating an image with a certain tag based on the annotation of another tag in a sample image set, and adds this probability value to the tag information of another image set to improve annotation efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual annotation is used for multi-tag images, then annotation accuracy can be maintained, but the consumption of manpower and material resources increases drastically

Engineering Contradiction:
Improveannotation accuracyVSAvoidannotation efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent introduces a computer system as an intermediary between the annotator and the annotation task. The system automatically performs tag propagation by calculating co-occurrence probabilities between tags and applying them to annotate images, reducing manual effort while maintaining annotation quality through algorithmic precision

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces the mechanical manual annotation process with an automated computational system. Instead of manually examining and annotating each image, the system uses co-occurrence analysis algorithms to automatically propagate tags across images, substituting human labor with computational processing

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

2Productivity

If automatic tag propagation is used, then annotation efficiency is improved, but the complexity of the annotation process increases

Engineering Contradiction:
Improveannotation efficiencyVSAvoidannotation process complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent performs preliminary co-occurrence analysis to build a probability matrix before actual tag propagation. By pre-calculating the relationships between tags and storing them in a structured matrix, the system simplifies the subsequent annotation process and makes it more manageable despite the underlying complexity

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If co-occurrence analysis is performed on large image sets, then tag propagation accuracy improves, but the time and computational resources required increase

Engineering Contradiction:
Improvetag propagation accuracyVSAvoidcomputation time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent segments the large-scale co-occurrence analysis into manageable components by organizing tags into a structured matrix format. This segmentation allows the system to process and store co-occurrence probabilities in an organized manner, making the computation more efficient and scalable for large image sets

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3822859B1Annotation method and device, and storage medium
Publication Date: 2025.02.19 BEIJING XIAOMI INTELLIGENT TECH CO LTD
  • EP3822859B1 patent drawingFigure 1A~1B
  • EP3822859B1 patent drawingFigure 1C
  • EP3822859B1 patent drawingFigure 2

AI summary

An annotation method and device and a storage medium are provided. The annotation method includes operations as follows. A first probability value that a first sample image is annotated with an Nth tag when the first sample image is annotated with an Mth tag is determined based on first tag information of a first image set (S11). M and N are unequal and are positive integers. The first probability value is added to second tag information of a second sample image annotated with the Mth tag in a second image set (S12).