Distance Image Block Segmentation for Privacy-Preserving State Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing detection techniques using Time-of-Flight cameras struggle to protect privacy while accurately detecting the state of a target object, such as a human, from three-dimensional distance images, as they often expose personal information.
Innovation Solution
A detection system that divides pixels in a distance image into blocks based on distance relationships, calculates virtual areas or volumes for each block, and compares pixel ratios to detect the state of the target object, allowing for abnormality detection without exposing personal information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If a general camera is used to detect the state of a target object, then attribute information such as portraits can be acquired, but personal information such as privacy is exposed
Solution Approach 1:
The patent segments the distance image into multiple blocks based on distance relationships, and further segments each block into pixel groups. This segmentation allows state detection without needing to process or store detailed attribute information, thereby protecting privacy while maintaining detection capability.
Solution Approach 2:
The patent extracts only the essential state information from the distance image by comparing pixel ratios between blocks, rather than extracting or processing detailed attribute information. This extraction approach enables state detection while eliminating privacy-exposing data.
2Loss of information
If a distance image is used to protect privacy, then personal information is not exposed, but it is not easy to detect the state on the target object from the three-dimensional information
Solution Approach 1:
The patent divides the distance image into multiple blocks based on distance relationships, and further divides each block into pixel groups. This segmentation transforms the complex three-dimensional distance information into a structured format that is easier to analyze for state detection while maintaining privacy protection.
Solution Approach 2:
The patent changes the parameter representation from raw distance values to pixel ratios between blocks. This parameter transformation simplifies the detection process by converting complex three-dimensional information into comparable ratio metrics that indicate target object states.
3Measurement precision
If detailed attribute information is processed to detect state, then detection accuracy is improved, but information processing load increases
Solution Approach 1:
The patent extracts only the necessary state information by comparing pixel ratios between blocks, discarding detailed attribute information. This selective extraction reduces information processing load while maintaining sufficient accuracy for state detection.
Solution Approach 2:
Instead of processing detailed attribute information to derive state, the patent inverts the approach by using simplified distance-based pixel ratio comparisons to directly detect state. This inversion reduces computational complexity while achieving the detection goal.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables accurate and privacy-protecting state detection of a target object by reducing the need for attribute information and minimizing information processing load, improving detection speed and accuracy.
Implementation Method 1
A Time-of-Flight Camera (ToF camera) is a camera capable of irradiating a target object with light and measuring three-dimensional information (distance image) from the target object using an arrival time of reflected light.
Data Source
AI summary
A detection system includes an imaging unit configured to acquire a distance image indicating a distance to a target object, and a processing unit configured to divide pixels included in the distance image into blocks according to a distance relationship, and detects a state of the target object by comparing ratios of pixels included in the respective blocks.