Image Processing Apparatus Fixed Region Masking

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image processing techniques fail to reliably conceal fixed regions when a moving object mask is superimposed, leading to insufficient privacy protection in monitored areas.

Innovation Solution

An image processing apparatus and method that detects moving objects and superimposes a moving object mask on the detected objects, while also applying a fixed mask on predetermined regions in the image, ensuring the fixed region's concealment by overlaying the fixed mask on the moving object mask.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a moving object mask is superimposed on detected moving objects, then moving objects can be concealed, but the concealment of fixed regions becomes insufficient when moving objects are present in fixed regions

Engineering Contradiction:
Improveconcealment reliabilityVSAvoidfixed region concealment
Core Design Contradiction:
ReliabilityVSLoss of information

Solution Approach 1:

The patent segments the masking operation into two distinct stages: first applying a moving object mask to detected moving objects, then applying a fixed region mask to predetermined fixed regions. This segmentation ensures that fixed regions are processed independently and their concealment is not compromised by the presence of moving objects within them.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary action by first superimposing the moving object mask on detected moving objects before applying the fixed region mask. This sequence ensures that when the fixed region mask is subsequently applied, it completely conceals any moving objects within fixed regions, preventing any leakage of information.

Inventive Principle:
Principle #10Preliminary action

2Loss of information

If a fixed mask is applied to predetermined regions, then fixed regions are concealed, but moving objects within those regions may still be detected and notified

Engineering Contradiction:
Improvefixed region concealmentVSAvoidconcealment reliability
Core Design Contradiction:
Loss of informationVSReliability

Solution Approach 1:

The patent merges two masking operations into a unified process: the moving object mask and the fixed region mask are both applied to the same image, with the fixed region mask taking precedence in fixed regions. This combination ensures that fixed regions are completely concealed while maintaining the overall privacy protection strategy.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent applies different masking qualities to different regions: moving objects outside fixed regions receive moving object mask treatment, while fixed regions receive fixed region mask treatment that completely conceals their contents. This local differentiation ensures that fixed regions have superior concealment quality compared to other areas.

Inventive Principle:
Principle #3Local quality

3Difficulty of detecting and measuring

If moving object detection is performed in fixed regions, then moving objects can be detected, but privacy protection in fixed regions becomes insufficient

Engineering Contradiction:
Improvemoving object detectionVSAvoidprivacy exposure
Core Design Contradiction:
Difficulty of detecting and measuringVSObject-affected harmful factors

Solution Approach 1:

The patent extracts moving objects detected within fixed regions from the visible output by applying the fixed region mask after moving object detection. This extraction allows the system to detect moving objects for monitoring purposes while completely removing their visual representation in the output image, thereby protecting privacy while maintaining detection capability.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP3605468B1Image processing apparatus, image processing apparatus control method, and non-transitory computer-readable storage medium
Publication Date: 2025.01.01 CANON KK
  • EP3605468B1 patent drawingFigure 1~2B
  • EP3605468B1 patent drawingFigure 3A~3C
  • EP3605468B1 patent drawingFigure 4

AI summary

An image processing apparatus which extracts a region of a foreground object in a captured image obtained by an imaging unit, and controls a display unit to display an output image formed by superimposing, on a predetermined image, a first image formed by anonymizing the extracted region, and a second image formed by anonymizing a region designated by a user on the captured image, wherein an appearance of the second image is different from an appearance of the first image, a region of a foreground object is extracted from a region other than at least the designated region in the captured image, and even when the designated region contains the foreground object, the first image in the designated region is not displayed on the output image.