Image Processing Apparatus Human Detection Precision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing apparatuses face challenges in accurately detecting user proximity due to the influence of attire, shoes, and belongings on contour information, leading to reduced detection precision when determining human presence.
Innovation Solution
The apparatus employs a combination of camera units to capture images, a correlation calculating unit to extract movement information using CHLAC characteristics, and a movement information determining unit to compare this information with stored dictionaries, thereby improving detection precision and reducing errors caused by non-human objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If contour information from human sensors is used to detect human presence, then detection range is extended, but detection precision deteriorates due to influence from attire, shoes, and belongings
Solution Approach 1:
The patent segments the detection task into two parts: a first human sensor for wide-area detection and a second human sensor for precise verification. By dividing the detection function across multiple sensors with different ranges, the system achieves both extended detection coverage and high precision in confirming actual human presence while filtering out false positives from objects like bags or shoes.
Solution Approach 2:
The patent introduces an image processing apparatus as an intermediary that captures images of the detected moving object and analyzes whether it represents an actual human. This intermediary verification step mediates between the broad detection capability of human sensors and the need for precise human identification, using image analysis to confirm whether the detected object is truly a human before triggering power saving state cancellation.
2Speed
If the power saving state is canceled immediately upon detecting a moving object, then response speed is improved, but error detection increases due to false positives from non-human objects
Solution Approach 1:
The patent performs preliminary verification by capturing images and analyzing whether the moving object is an actual human before canceling the power saving state. This preliminary action filters out false positives from non-human objects while maintaining fast response for genuine human presence, as the image analysis is performed quickly using the already-captured image data.
Solution Approach 2:
The patent implements a feedback mechanism where the image processing apparatus continuously monitors and provides verification feedback on detected moving objects. The system uses the image analysis results to confirm or reject the detection, creating a closed-loop verification process that maintains high response speed while ensuring reliability by only canceling power saving state when human presence is confirmed.
3Measurement precision
If multiple human sensors with different detection ranges are used, then detection accuracy is improved, but device complexity increases
Solution Approach 1:
The patent makes the image processing apparatus multi-functional by combining wide-area detection capability with precise human verification functionality in a single integrated system. Rather than requiring separate dedicated sensors for each function, the image processing apparatus performs both the initial detection and the verification tasks, reducing overall system complexity while maintaining high detection accuracy.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
There is provided an image processing apparatus comprising: a capturing part configured to sequentially capture a plurality of images; an extracting part configured to extract movement information from the sequentially captured images through an analysis, wherein the movement information relates to movement of a moving object and the sequentially captured images contain images of the moving object; and a first determining part configured to detect an approaching user based on the extracted movement information and approaching movement information, stored in advance, indicating approaching movement of a human.