Visual Query System for Textual Term Identification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Users often face difficulties in locating information related to objects they see in the physical world without knowing the appropriate search terms for term-based queries, as existing systems require explicit input of words or phrases.

Innovation Solution

A visual query system that processes images to determine corresponding textual terms by utilizing a server network with parallel search systems, including image-to-terms search, facial recognition, and OCR, to generate relevant search results.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If term-based searching is used, then search results can be obtained, but users must explicitly provide search terms which they may not know

Engineering Contradiction:
Improvesearch accuracyVSAvoiduser convenience
Core Design Contradiction:
Measurement precisionVSEase of operation

Solution Approach 1:

The patent introduces an image-to-terms search system as an intermediary between the user's visual query and the term-based search engine. The system automatically converts images into search terms through multiple parallel search systems (OCR, facial recognition, image-to-terms), allowing users to search without knowing appropriate keywords while maintaining search accuracy

Inventive Principle:
Principle #24Intermediary (Mediator)

2Adaptability or versatility

If multiple parallel search systems are implemented, then comprehensive search capabilities are achieved, but system complexity increases

Engineering Contradiction:
Improvesearch capabilityVSAvoidsystem structure
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent divides the search system into multiple independent parallel search systems, each handling specific tasks (OCR for text extraction, facial recognition for face identification, image-to-terms for object recognition). This segmentation allows comprehensive search capabilities while managing complexity through modular, specialized components rather than a monolithic system

Inventive Principle:
Principle #1Segmentation

3Ease of operation

If image-to-terms mapping is performed, then visual queries can be converted to search terms, but processing time increases

Engineering Contradiction:
Improvesearch accessibilityVSAvoidprocessing time
Core Design Contradiction:
Ease of operationVSLoss of time

Solution Approach 1:

The patent performs preliminary actions by pre-processing images through multiple parallel search systems simultaneously upon receiving a visual query. The system extracts text, identifies faces, and generates object terms in parallel before ranking and presenting results, reducing overall processing time compared to sequential processing

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP2883158B1Identifying textual terms in response to a visual query
Publication Date: 2019.11.06 GOOGLE LLC
  • EP2883158B1 patent drawingFigure 1
  • EP2883158B1 patent drawingFigure 2
  • EP2883158B1 patent drawingFigure 3

AI summary

A method, system, and computer readable storage medium is provided for identifying textual terms in response to a visual query is provided. A server system receives a visual query from a client system. The visual query is responded to as follows. A set of image feature values for the visual query is generated. The set of image feature values is mapped to a plurality of textual terms, including a weight for each of the textual terms in the plurality of textual tenns. The textual terms are ranked in accordance with the weights of the textual terms. Then, in accordance with the ranking the textual tenns, one or more of the ranked textual tenns are sent to the client system.