Image Search Error Calibration via User Feedback

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image search methods fail to accurately reflect user intentions due to differences in preferences among users and errors between selected words and intended images, leading to unsatisfied user needs.

Innovation Solution

A method that assigns prime and style keywords to images, calculates confidence intervals, and calibrates errors by collecting data from users to build a database, allowing for the display of target images that match user preferences by extracting images within a specific error range and style similarity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If image search is performed using automatically generated candidate images and user feedback, then user preference reflection is improved, but error between user intention and selected images cannot be fully eliminated due to individual differences

Engineering Contradiction:

Improveaccuracy of user intention matchingVSAvoidindividual user preference variation

Core Design Contradiction:

Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The system performs preliminary error calibration by presenting test images and collecting user feedback before actual image search operations. This preliminary action establishes individual user error patterns that are then applied to correct future search results, improving accuracy while accounting for individual preferences

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system implements continuous feedback loops where user selections and corrections are collected during actual use, and this feedback is used to refine the error calibration models. This ongoing feedback mechanism allows the system to adapt to individual user preferences while maintaining high matching accuracy

Inventive Principle:

Principle #23Feedback

2Productivity

If keywords are assigned to images based on general user data, then database building is efficient, but error calibration for individual users is insufficient

Engineering Contradiction:

Improvedatabase building efficiencyVSAvoidindividual error calibration accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system segments the error calibration process into two parts: a general database building phase using aggregated user data for efficiency, and an individual calibration phase that applies personalization to each user. This segmentation allows both efficient database construction and precise individual error calibration

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system performs preliminary error calibration using test images before actual search operations begin. This preliminary action establishes baseline error patterns for each user without requiring extensive data collection during productive search operations, maintaining both efficiency and precision

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS10318577B2Method for identifying user's needs based on image word interpretation

Publication Date: 2019.06.11 PERCEPTION INC

AI summary

A method for identifying a need of a user based on an image word interpretation, the method comprising: assigning prime keywords and style keywords to a plurality of images, wherein the images having the prime and style keywords assigned thereto are stored by at least one image category of the images; receiving a target keyword from the user depending on a need of a user; displaying images within the at least one image category, and receiving a preferred image by the at least one category from the user; extracting a target image that matches the target keyword based on the received preferred image; and displaying the extracted target image in accordance with at least one way.