Image Sensing Apparatus for Low-Power User Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing user detection methods based on image recognition require high-quality cameras, significant computational power, and continuous analysis of pixel relations, making it difficult to provide active services efficiently due to high power consumption and complex computations.
Innovation Solution
An image sensing apparatus using a low-power image sensor that analyzes images by obtaining first and second difference images between current and previous frames, and background images, allowing for efficient user existence and position determination with minimal computation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If object-recognition algorithm is used to detect user from images, then user detection accuracy is improved, but computation power requirement increases significantly
Solution Approach 1:
The patent segments the image analysis process into two distinct difference calculation stages: first calculating difference between current and previous frames to detect motion, then calculating difference between current frame and background to identify static objects. This segmentation avoids the need for complex object-recognition algorithms while maintaining effective user detection capability.
Solution Approach 2:
The patent extracts only the essential motion and presence information from images by calculating simple pixel difference values between frames and background, discarding the need for complex feature extraction and recognition. This extraction approach maintains detection effectiveness while dramatically reducing computational requirements.
2Measurement precision
If continuous image analysis is performed to track user position, then user position tracking accuracy is improved, but power consumption increases
Solution Approach 1:
The patent performs image analysis at periodic intervals by comparing current frames with previously stored frames and background images, rather than continuously analyzing every pixel relationship. This periodic difference calculation maintains position tracking accuracy while significantly reducing power consumption compared to continuous real-time analysis.
3Reliability
If high-quality camera is used for user detection, then detection reliability is improved, but device cost and complexity increase
Solution Approach 1:
The patent creates difference images as simplified copies of the original image data, extracting only the essential motion and presence information through pixel comparison. This copying approach allows reliable user detection using simple image sensors without requiring high-quality cameras, as the difference calculation process compensates for lower image quality.
4Speed
If object-chase algorithm is applied in addition to object-recognition, then real-time tracking is improved, but computation complexity increases
Solution Approach 1:
The patent merges the object detection and tracking functions into a single unified process by continuously calculating difference images between current frames and reference frames. This combination eliminates the need for separate object-recognition and object-chase algorithms, achieving real-time tracking with reduced computational complexity.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
An apparatus and a method of detecting an object of the apparatus are provided. The apparatus includes a sensing part configured to photograph an image, a storage configured to store a background image frame, and a controller configured to obtain a first difference image and a second difference image from the photographed image and to determine an existence and a position of an object using the first and the second difference images, wherein the first difference image is an image indicating a difference between a currently photographed image frame and a previously photographed image frame, and the second difference image is an image indicating a difference between the currently photographed image frame and the background image frame stored in the storage.