Image Processing Apparatus Mixed Voice Character String Combination

Solution Overview

Problem

Existing techniques for combining character strings with still or moving images fail to produce meaningful results when the voice within the moving image is a mixed voice, leading to unnatural character strings being generated.

Innovation Solution

An image processing apparatus that selects parts of a moving image, extracts voices during specific times, and combines character strings based on predetermined conditions, ensuring that only appropriate character strings are used when the voice is mixed, using a combination of selection, extraction, and processing units to handle mixed voices effectively.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a character string corresponding to the voice is combined in a moving image, then the image processing function is improved, but unnatural character strings are generated when the voice is a mixed voice

Engineering Contradiction:

Improveimage processing functionVSAvoidcharacter string accuracy

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The system performs preliminary voice quality assessment before combining character strings with images. By evaluating whether the voice is a mixed voice in advance, the system determines whether to proceed with automatic character string combination or to use predetermined character strings, thereby preventing unnatural combinations from occurring in the first place

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system introduces an intermediate evaluation step that acts as a mediator between voice extraction and character string combination. This intermediary process assesses voice quality and controls the flow to appropriate processing paths, ensuring that character strings are only combined when the voice is suitable

Inventive Principle:

Principle #24Intermediary (Mediator)

2Productivity

If automatic character string combination is performed, then processing efficiency is improved, but meaningless character strings are generated from mixed voices

Engineering Contradiction:

Improveprocessing efficiencyVSAvoidvoice meaning accuracy

Core Design Contradiction:

ProductivityVSLoss of information

Solution Approach 1:

The system performs preliminary voice quality assessment before combining character strings with images. By evaluating whether the voice is a mixed voice in advance, the system determines whether to proceed with automatic character string combination or to use predetermined character strings, thereby preventing unnatural combinations from occurring in the first place

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system introduces an intermediate evaluation step that acts as a mediator between voice extraction and character string combination. This intermediary process assesses voice quality and controls the flow to appropriate processing paths, ensuring that character strings are only combined when the voice is suitable

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS11394919B2Image processing apparatus, image processing method and medium

Publication Date: 2022.07.19 CANON KK

AI summary

An object of one embodiment of the present disclosure is to provide a product with a high added value to a user by preventing an erroneous character string from being combined in a case where a voice before and after an image selected from within a moving image is a mixed voice. One embodiment of the present disclosure is an image processing apparatus including: a selection unit configured to select, from a moving image including a plurality of frames, a part of the moving image; an extraction unit configured to extract a voice during a predetermined time corresponding to the selected part in the moving image; and a combination unit configured to combine a character string based on a voice extracted by the extraction unit, with the part of the moving image selected by the selection unit or a frame among frames corresponding to the part.