Information Processing Apparatus for Accurate Character Extraction Guidance

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In scenarios where an image contains multiple pieces of search information, it is unclear which character information to extract as a value, leading to degraded accuracy in extraction processes.

Innovation Solution

An information processing apparatus that receives type information about a target and outputs guidance to acquire a first image including search information and its vicinity, using a processor to guide the user in photographing, enlarging, or marking the relevant regions for accurate extraction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If key-value extraction is performed on images containing multiple search information pieces, then character information can be extracted, but accuracy in extraction is degraded because it is unclear near which search information to acquire the value

Engineering Contradiction:

Improveextraction accuracyVSAvoidambiguity in search information identification

Core Design Contradiction:

Measurement precisionVSLoss of information

Solution Approach 1:

The patent segments the image processing task by first identifying and extracting individual search information pieces from the image, then performing key-value extraction for each segmented search information separately. This segmentation eliminates the ambiguity of which search information to associate with a value by treating each search information as an independent extraction target.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary action by first detecting and locating all search information pieces in the image before performing the actual key-value extraction. By pre-identifying the positions and contents of search information, the system prepares the extraction process in advance, ensuring that each value is correctly associated with its corresponding search information without ambiguity.

Inventive Principle:

Principle #10Preliminary action

2Reliability

If OCR is performed on the entire image, then all text can be recognized, but processing time and computational resources increase significantly

Engineering Contradiction:

Improvetext recognition completenessVSAvoidprocessing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent extracts only the necessary regions containing search information and their associated values from the entire image, rather than performing OCR on the whole image. By extracting and processing only the relevant portions, the system maintains complete text recognition for the needed information while significantly reducing processing time and computational resources.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies local quality by performing high-precision OCR only in the specific regions where search information and its corresponding values are located, rather than uniformly processing the entire image. This localized approach ensures reliable text recognition where needed while avoiding unnecessary processing in other areas, thus reducing overall processing time.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS11462014B2Information processing apparatus and non-transitory computer readable medium

Publication Date: 2022.10.04 FUJIFILM BUSINESS INNOVATION CORP

AI summary

An information processing apparatus includes a processor configured to, when extracting character information for search information corresponding to a type of a target from an image obtained by photographing the target, receive type information indicating the type of the target, and output first guidance information for providing a guidance to acquire a first image including at least search information associated with the type information and a vicinity of the search information.