Dynamic Stand-up Detection Line for Bed Leaving
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing bed leaving detection systems in medical and nursing care facilities face false detection issues due to camera installation constraints and motions other than standing up, such as rolling over or sitting, which can lead to inaccurate alerts.
Innovation Solution
A behavior detection method that dynamically sets a stand-up detection line based on the predicted movement path of a patient's head from a forward bending posture, allowing for accurate differentiation between standing up and other motions by setting the line in a position that does not overlap with forward bending detection lines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a camera is installed in an oblique direction to the bed to avoid interfering with treatment or caregiving, then the ease of operation is improved, but the measurement precision deteriorates because the detection lines overlap with the bed area causing false detection
Solution Approach 1:
The patent applies dynamics by making the stand-up detection line movable rather than fixed. The line is dynamically adjusted based on the detected head position and predicted movement path, allowing it to adapt to different camera angles and patient motions. This resolves the contradiction by enabling oblique camera installation while maintaining detection accuracy through real-time line repositioning.
Solution Approach 2:
The patent changes the parameter of the detection line position from a fixed value to a dynamically variable value. By calculating the head movement path and predicting the stand-up trajectory, the system adjusts the detection line parameters (position and orientation) to match the actual motion pattern, thereby maintaining measurement precision regardless of camera installation angle.
2Device complexity
If fixed detection lines are used to determine bed leaving, then the device complexity is reduced, but the reliability deteriorates due to false detection from motions other than standing up
Solution Approach 1:
The patent applies preliminary action by detecting the forward bending motion before the stand-up action occurs. The system first identifies when the patient bends forward, then predicts the subsequent head movement path, and finally sets the stand-up detection line accordingly. This sequential approach improves reliability by distinguishing standing up from other motions while adding only minimal processing steps.
Solution Approach 2:
The patent implements feedback by using the detected head position and movement path to continuously adjust the stand-up detection line. The system monitors the patient's motion, predicts the stand-up trajectory based on observed behavior, and recalibrates the detection line in real-time, creating a closed-loop system that improves reliability without significant complexity increase.
Data Source
AI summary
A behavior detection apparatus detects a position of a head from an image; based on the position of the head, detects a motion of bending forward when a person to be detected sits up in a width direction of a bed in a bed area of the image and stands up from the sitting state; when a movement of the head is detected to be forward bending, sets, based on a movement path of the head, a stand-up detection line for determining bed leaving of the person between a lowest point of the forward bending and the position of the head when the person stands by the bed, the stand-up detection line being set in a position not overlapping with a forward bending detection line by which the motion of bending forward is detected; and detects a stand-up motion when the head passes the set stand-up detection line.