TWI744990B - Care system for predicting bed exit and patient management system including the same - Google Patents

Abstract

A care system for predicting bed exit is provided, in which is suitable for a frame for supporting a patient. The care system includes at least one boundary-crossing detection system configured to be coupled with the frame, a location sensing system configured to be coupled with the frame, and a control circuit. The location sensing system is configured to obtain relative position information between the patient and the frame. The control circuit is configured to perform a data communication with the at least one boundary-crossing detection system and the location sensing system, and also configured to access care date configured to define multiple bed-exit behaviors performed by the patient on the frame. When at least one of the following situations occurs, the control circuit transmits a warning message: (1) the control circuit determines that, according to the relative position information and the care date, a current behavior of the patient corresponds to one of the multiple bed-exit behaviors; (2) the at least one boundary-crossing detection system detects that an object is passing through.

Description

Nursing system and patient management system for predicting the behavior of getting out of bed

This disclosure is related to a care system and patient management system, especially a care system and patient management system for predicting the behavior of getting out of bed.

As Taiwan’s population is aging, the demand for medical care for the elderly has also increased. Most elderly people have osteoporosis, chronic diseases, etc., so when they fall, they often cause serious sequelae. In view of the fact that the cause of elderly falls is usually due to their own getting in and out of bed, the use of bed-out alarms in medical institutions can reduce the occurrence of falls and improve the safety of the elderly in hospital.

The present disclosure provides a care system for predicting the behavior of getting out of bed, which is suitable for a frame for supporting patients. The care system includes at least one cross-border detection system coupled to the frame, a position sensing system coupled to the frame, and a control circuit. The position sensing system is used to obtain the relative position information between the patient and the frame. The control circuit is used for detecting at least one out-of-bounds The system communicates with the position sensing system, and is used to access the care data used to define the patient's multiple out-of-bed behaviors on the frame. When at least one of the following situations occurs, the control circuit sends a warning message: (1) The control circuit determines that the patient’s current behavior corresponds to one of multiple out-of-bed behaviors based on the relative position information and the care data; (2) At least one out of range behavior The detection system senses the passage of the object.

The present disclosure provides a patient management system, which includes one or more care systems and main control devices. Each care system is suitable for a frame for supporting patients, and includes at least one cross-border detection system for coupling to the frame, a position sensing system for coupling to the frame, and a control circuit. The position sensing system is used to obtain the relative position information between the patient and the frame. The control circuit is used for data communication with at least one out-of-bounds detection system and position sensing system, and is used for accessing care data used to define multiple out-of-bed behaviors of patients on the frame. When at least one of the following situations occurs, the control circuit sends a warning message: (1) The control circuit determines that the patient’s current behavior corresponds to one of multiple out-of-bed behaviors based on the relative position information and the care data; (2) At least one out of range behavior The detection system senses the passage of the object. The main control device is used for data communication with the control circuit to receive and correspondingly display warning messages, and is used to provide a main control terminal user interface. The host user interface includes one or more sensitivity adjustment patterns respectively corresponding to one or more care systems. Each sensitivity adjustment pattern is a user-interactive object. When the sensitivity adjustment pattern is moved from the first preset area to the second preset area, the control circuit corresponding to the sensitivity adjustment pattern in one or more care systems will change one of the multiple out-of-bed behaviors to that of multiple out-of-bed behaviors. The other.

One of the advantages of the above multiple embodiments is that it can avoid system failure and judgment delay caused by network congestion, failure, or insufficient bandwidth.

Another advantage of the above embodiments is that the patient's getting out of bed can be detected earlier, so that more time can be obtained for the frontline nursing staff to deal with the out of bed event.

Another advantage of the above-mentioned multiple embodiments is to provide diversified channels and platforms for obtaining information, so that medical staff can grasp the patient's out-of-bed alarm in real time.

100,600: care system

101: Frame

103a, 103b, 103c, 103d: armrest

105a, 105b: plate-shaped member

107: Edge

110: Cross-border detection system

111: The first distance sensor

112: The second distance sensor

120A, 120B: position sensing system

122: The first position sensor

124: second position sensor

130: control circuit

132: Care Information

140: Audio and video capture device

150: first sensing area

160: second sensing area

EV: Event notification

Da-1~Da-n: first distance

Db-1~Db-n: second distance

ODa-1~ODa-n: the first optimized distance

ODb-1~ODb-n: the second optimal distance

D1: First direction

D2: second direction

210-1~210-n: Distance sensing unit

220, 310, 720: Microprocessor

230: Three-axis accelerator

240, 320, 730: communication interface

330: Voice Module

510,520: Curve

Za, Zb, Zc: zero crossing point

Aa, Ab: area

P1, P2, P3, P4, P5: coordinate points

La,Lb,Lc: Connect

710: Thermal imaging circuit

800A, 800B, 800C: method

S802A~S810A, S802B~S808B, S802C~S810C: Process

910: Image

920: Foreground image

10: Patient Management System

12: Master control device

14,14-1,14-2,14-3: care system

16: Cloud server

18: mobile device

20: Console user interface

22, 22-1, 22-2, 22-3: display area

SP: Sensitivity adjustment pattern

30: Field of view

40: Mobile user interface

42: Geographical information

44: Medical History Information

46: Live image area

Figure 1 is a schematic diagram of a care system according to an embodiment of the present disclosure.

FIG. 2 is a simplified functional block diagram of the first distance sensor according to an embodiment of the present disclosure.

FIG. 3 is a simplified functional block diagram of the position sensing system according to an embodiment of the present disclosure.

Figure 4A is a broken line diagram exemplarily drawn from the original data of the distance between the patient and the position sensing system when the patient passes in front of the position sensing system.

FIG. 4B is a broken line graph exemplarily drawn based on the first optimized distance and the second optimized distance.

Figure 5A exemplarily shows the zero crossing point of the curve formed by subtracting the first optimized distance and the second optimized distance.

Figure 5B exemplarily shows the area above and below the coordinate axis of the curve in Figure 5A.

Fig. 5C exemplarily shows the line of coordinate points of the curve in Fig. 5A.

Figure 5D exemplarily shows the line of coordinate points of another curve.

FIG. 5E exemplarily shows the distribution relationship between the first optimized distance and the second optimized distance on the time axis.

Figure 6 is a schematic diagram of a care system according to an embodiment of the present disclosure.

FIG. 7 is a simplified functional block diagram of the position sensing system according to an embodiment of the present disclosure.

Figure 8A is a flow chart of the probability of generating a behavior corresponding to a single image.

Figure 8B is a flowchart showing the probability of actions corresponding to multiple images.

Figure 8C is a flow chart of generating behavior probabilities corresponding to multiple divided images.

FIG. 9A exemplarily shows one of the multiple images captured by the position sensing system.

Fig. 9B exemplarily shows the foreground image obtained after the image of Fig. 9A has been de-backed.

Figure 10 is a simplified functional block diagram of the patient management system according to an embodiment of the present disclosure.

FIG. 11 is a schematic diagram of a host user interface according to an embodiment of the present disclosure.

Fig. 12 illustrates a schematic diagram of the field of view of the user of the mobile device.

The embodiments of the present disclosure will be described below in conjunction with related drawings. In the diagrams, the same reference numerals indicate the same or similar components or method flows Procedure.

FIG. 1 is a schematic diagram of a care system 100 according to an embodiment of the present disclosure. The components of the care system 100 can be all arranged on a frame 101 for supporting the human body, or partly arranged on the frame 101 and partly arranged on the environment around the frame 101 (for example, a wall). The frame 101 includes a plurality of armrests 103a~103d that can be raised and lowered, and includes plate-shaped members 105a~105b respectively located at both ends of the frame 101, wherein the armrests 103a~103d are arranged on the frame 101 parallel to the first direction D1. There are several edges (for example, the edge 107), and the plate-shaped members 105a-105b are arranged on a plurality of edges of the frame 101 parallel to the second direction D2. In some embodiments, the frame 101 is a medical bed used to support a patient and a mattress. The care system 100 can proactively notify the medical staff to assist when detecting the patient’s getting out of bed, so as to reduce fall injuries.

As shown in Figure 1, the care system 100 includes a boundary-crossing detection circuit 110, a position sensing system 120A, a control circuit 130, and an audio-visual capture device 140. The control circuit 130 is communicatively connected to the boundary-crossing detection circuit. The system 110, the position sensing system 120A, and the video and audio capture device 140. The cross-border detection system 110 is partly arranged on the armrest 103a, and partly arranged on the end of the edge 107 connected to the armrest 103a near the plate-shaped member 105b to detect whether there is an object (such as the patient's limbs) passing through the plate-shaped member 105b and the armrest The space between 103a.

In this embodiment, the cross-border detection system 110 includes a first distance sensor 111 and a second distance sensor 112, wherein the first distance sensor 111 is arranged on the side of the armrest 103a close to the plate-shaped member 105b, and the second distance sensor 112 is provided with an edge 107 and is located between the armrest 103a and the plate-shaped member 105b. The first distance sensor 111 and the second distance sensor 112 have a first sensing area 150 and a second sensing area 160, respectively.

In some embodiments, when the armrest 103a is in the upright state, the first sensing area 150 and the second sensing area 160 at least partially overlap in space. When an object passes through the first sensing area 150 and/or the second sensing area 160, the out-of-bounds detection system 110 will send an event notification EV to the control circuit 130. When the control circuit 130 receives the event notification EV, the control circuit 130 can transmit the warning message via the network via wired or wireless communication. For example, the control circuit 130 can send a warning message to the nursing station main control device 12, the cloud server 16 and the mobile device 18 of the medical staff in Figure 10 described later.

The situation in which the aforementioned object passes may be, for example, that part of the object is located within the first sensing area 150 and/or the second sensing area 160 (for example, the patient is sitting on the edge of the bed), or the part of the object is located in the first sensing area 150. And/or move in the second sensing area 160 (for example, the patient is extending his hands and feet out of the edge of the bed).

In some embodiments, the second distance sensor 112 may also be disposed on the side of the plate-shaped member 105b close to the first distance sensor 111, that is, the first distance sensor 111 and the second distance sensor 112 is substantially parallel and oppositely arranged. At this time, when the armrest 103a is in the upright state, the first sensing area 150 and the second sensing area 160 will also be at least in the space. Partially overlapped.

In other embodiments, the care system 100 may include one or more out-of-bounds detection systems 110. These out-of-bounds detection systems 110 may be installed on the armrests 103a~103d in a manner similar to the above, and the frame 101 is parallel to the first direction. The multiple edges of D1 are located between the armrests 103a-103d and the corresponding plate-shaped members 105a-105b, or on the plate-shaped members 105a-105b. In this way, the care system 100 can detect the patient's getting out of bed in multiple directions.

FIG. 2 is a simplified functional block diagram of the first distance sensor 111 according to an embodiment of the present disclosure. The first distance sensor 111 includes a plurality of distance sensing units 210-1 to 210-n, a microprocessor 220, a three-axis accelerator 230, a communication interface 240, and a power module (not shown), of which the microprocessor 220 It is coupled to each of the other functional blocks in FIG. 2, and the first distance sensor 111 is communicatively connected to the control circuit 130 through the communication interface 240. In some embodiments, the distance sensing units 210-1 to 210-n each include one or more infrared transceiving circuits to form the first sensing area 150 by using multiple infrared light beams. The microprocessor 220 is used to determine whether an object passes through the first sensing area 150, and is used to send an event notification EV to the control circuit 130 through the communication interface 240. In order to make the drawing concise and easy to explain, the other components and the connection relationship in the first distance sensor 111 are not shown in the second figure.

The three-axis accelerator 230 is used to transmit the magnitude and/or direction of the acceleration of the armrest 103a to the microprocessor 220, so that the microprocessor 220 can determine the placement state of the armrest 103a. When the microprocessor 220 judges that the armrest 103a When there is a substantially downward acceleration, the microprocessor 220 will determine that the armrest 103a is switched from the upright state to the down state. At this time, the microprocessor 220 also sends an event notification EV to the control circuit 130 through the communication interface 240.

The aforementioned connection methods, components, implementations, and advantages of the first distance sensor 111 are also applicable to the second distance sensor 112, that is, the second sensing area 160 is also formed by multiple infrared light beams. For the sake of brevity, the corresponding content will not be repeated here. In some embodiments, the three-axis accelerator 230 of the first distance sensor 111 and/or the second distance sensor 112 may be omitted.

Please refer to Figure 1 again. The position sensing system 120A is set on one of the armrests 103a~103d to obtain the relative position information between the patient on the frame 101 and the frame 101, that is, to obtain the patient on the frame 101 Related information such as the body position or movement trend on the computer. Please refer to FIG. 1 and FIG. 3 at the same time. The position sensing system 120A includes a first position sensor 122 and a second position sensor 124. The sensing direction of the first position sensor 122 is substantially aligned with the upper half of the frame 101 (for example, the head of the bed to the center of the bed), and is used to sense the position of the patient's body on the upper half of the frame 101 multiple times. A plurality of first distances Da-1 to Da-n between the patient and the first position sensor 122 are obtained. The sensing direction of the second position sensor 124 is substantially aligned with the lower half of the frame 101 (for example, the foot of the bed to the center of the bed), and is used to sense the position of the patient's body in the lower half of the frame 101 multiple times to obtain A plurality of second distances Db-1 to Db-n between the patient and the second position sensor 124. The aforementioned relative position information includes these first distances Da-1~Da-n and The second distance Db-1~Db-n.

In some embodiments, the sensing ranges of the first position sensor 122 and the second position sensor 124 at least partially overlap. In practice, the first position sensor 122 and the second position sensor 124 can be implemented by ultrasonic ranging modules, respectively.

As shown in Figure 3, the position sensing system 120A also includes a microprocessor 310, a communication interface 320, and a voice module 330. The microprocessor 310 is coupled to each of the other functional blocks in Figure 3, and the position The sensing system 120A is communicatively connected to the control circuit 130 through the communication interface 320. The microprocessor 310 transmits the relative position information including the first distance Da-1~Da-n and the second distance Db-1~Db-n to the control circuit 130 through the communication interface 320, and the control circuit 130 will respond according to the relative position information. The location information is calculated with the care data 132 stored in advance to determine whether the patient's current behavior corresponds to one of the preset multiple out-of-bed behaviors. In some embodiments, various behaviors of getting out of bed can be, for example, lowering the armrests 103a~103d, moving from the head of the bed (for example, the plate-shaped member 105a) to the foot of the bed (for example, the plate-shaped member 105b), sitting on the edge of the bed, and getting out of the bed completely. The judgment process will be detailed in subsequent paragraphs. If the control circuit 130 determines that the patient is getting out of the bed, the control circuit 130 can send a warning message via the network via wired or wireless communication. For example, the control circuit 130 can send a warning message to the main control device 12, the cloud server 16 and the mobile device 18 of the medical staff in Figure 10 described later.

In some embodiments, when the control circuit 130 transmits a warning message, the control circuit 130 also instructs the voice of the position sensing system 120A The module 330 plays a preset voice notification to remind the patient to stop the behavior of getting out of bed. At this time, the control circuit 130 can also activate the audio-visual capture device 140 to transmit real-time images and sounds of the patient, for example, to transmit them to the mobile device 18 in FIG. 10 described later. In other embodiments, the nursing staff can communicate with the patient via video and/or audio via the mobile device 18 and the audio-visual capture device 140 to discourage the patient from getting out of bed.

It is worth noting that the control circuit 130 in Figure 1 can be installed independently, or can be installed inside the package housing of the cross-border detection system 110 or the position sensing system 120A. In some embodiments, the control circuit 130 can be integrated with the microprocessor 220 of FIG. 2 into a single chip, and the communication interface 240 is used to communicate with the position sensing system 120A and the audio-visual capture device 140 at this time. In other embodiments, the control circuit 130 can be integrated with the microprocessor 310 of FIG. 3 into a single chip. In this case, the communication interface 320 is used to communicate with the cross-border detection system 110 and the audio-visual capture device 140. That is, the various settings and changes made to the control circuit 130 according to the above teachings are all within the scope of this project.

The communication interfaces of various embodiments of the present disclosure (for example, the communication interfaces 240 and 320 in FIGS. 2 and 3) may be compatible with Bluetooth, Wi-Fi, ZigBee, cellular radio or any other wireless The communication technology of the communication protocol is implemented by hardware and/or software, but this disclosure is not limited to this. In some embodiments, the communication interface of the present disclosure may also be a communication interface using wired communication technology.

In addition, the control circuit 130 and the microprocessor (such as the microprocessor 220 and the microprocessor 310) of the various embodiments of the present disclosure may each be It can be realized by single-chip or multi-chip general-purpose processors, digital signal processors (DSP), field programmable logic gate arrays (FPGA), other suitable programmable devices, or a combination of one or more of the above.

The following will cooperate with Figures 4A~4B and Figures 5A~5E to illustrate that the control circuit 130 uses the first distance Da-1~Da-n and the second distance Db-1~Db-n to determine whether the patient has the behavior of getting out of bed. Related processes.

For example, Figure 4A shows the difference between the first distance Da-1~Da-n and the second distance Db-1~Db-n detected by the position sensor system 120A when the patient passes in front of the position sensor system 120A. The broken line diagram exemplarily drawn in the original data, for example, the patient gets up from the head of the bed (the plate-shaped member 105a) and moves to the foot of the bed (the plate-shaped member 105b). Since the ultrasound may be absorbed by the patient’s quilt or clothing, the distance values at multiple time points in this line chart suddenly return to zero, which makes the line chart have multiple data missing, which is not conducive to subsequent analysis. Therefore, the control circuit 130 is configured to perform a first optimization operation and a second optimization operation multiple times to respectively generate a plurality of first optimized distances ODa-1~ODa-n and a plurality of second optimized distances ODb-1~ ODb-n to compensate for missing data.

In detail, the first optimization operation is: selecting consecutive M first distances among the first distances Da-1~Da-n, and then selecting one of the M first distances with predetermined characteristics as the first optimization One of the distance between ODa-1~ODa-n. The second optimization operation is: select consecutive M second distances among the second distances Db-1~Db-n, and then select one of the M second distances with predetermined characteristics as the second optimized distance ODb-1 ~One of ODb-n. In some embodiments, the above-mentioned predetermined feature may be the largest among the M first distances in the first optimization operation. The numerical value, the average value or the smallest value of the M first distances, and in the second optimization operation, it can be the largest value of the M second distances, the average value of the M second distances, or the smallest value By.

FIG. 4B is a line graph exemplarily drawn based on the first optimized distance ODa-1~ODa-n and the second optimized distance ODb-1~ODb-n. It can be seen from Figure 4B that the above-mentioned first optimization operation and second optimization operation successfully fill in the missing values, effectively reducing the error values caused by the absorption of ultrasound waves by the object, which is conducive to the follow-up of the current behavior of the relevant patients. analyze. Next, the control circuit 130 subtracts the first optimized distance ODa-1~ODa-n from the second optimized distance ODb-1~ODb-n to obtain multiple coordinate points on the time axis, and then calculates the first coordinate point based on these coordinate points. The various characteristics shown in Figures 5A~5D are used to judge the patient’s current behavior based on these characteristics.

In this embodiment, since the patient moves from the head of the bed to the end of the bed, the first optimal distance ODa-1~ODa-n is positive in the first half of the entire time axis, and approaches in the second half of the entire time axis. Yu zero. Therefore, the curve 510 formed by the multiple coordinate points obtained by subtracting the first optimized distance ODa-1~ODa-n and the second optimized distance ODb-1~ODb-n in Figures 5A~5C is on the time axis The part in the first half is positive, and the part in the second half of the time axis is negative.

Please refer to FIG. 5A first, the control circuit 130 will calculate the number of zero-crossing points of the curve 510, that is, the number of times the curve 510 intersects the time axis. For example, the curve 510 has a zero crossing point Za. The number of zero crossing points can be used to determine the number of times the patient has passed in front of the position sensing system 120A.

Next, referring to FIG. 5B, the control circuit 130 calculates the ratio of the area of the curve 510 above the time axis to the area below the time axis. For example, the control circuit 130 calculates the ratio of the area Aa to the area Ab. This area ratio can be used to prevent the patient's minor movements from being misjudged as getting out of bed.

Please refer to FIG. 5C. The control circuit 130 selects two coordinate points from both sides of each zero-crossing point, for example, the coordinate point P1 and the coordinate point P2 located on the left and right sides of the zero-crossing point Za, respectively. Then, the control circuit 130 calculates the sum of the slopes of all the lines formed by the two coordinate points on both sides of each zero-crossing point. For example, the control circuit 130 calculates the slope of the line La formed by the coordinate point P1 and the coordinate point P2. The sum of the slopes can be used to determine the speed and direction of the patient's body. FIG. 5D exemplarily shows a curve 520 formed when the patient passes back and forth in front of the position sensing system 120A. The curve 520 has a zero-crossing point Zb and a zero-crossing point Zc. In order to calculate the total slope, the control circuit 130 selects the coordinate points P3 and P4 located on both sides of the zero crossing point Zb, and selects the coordinate points P4 and P5 located on both sides of the zero crossing point Zc. Next, the control circuit 130 calculates the sum of the slopes of the line Lb formed by the coordinate point P3 and the coordinate point P4 and the line Lc formed by the coordinate point P4 and the coordinate point P5. At this time, since the sum of the slopes is approximately zero, the control circuit 130 can classify the data in the 5D graph as false touches or noise.

Please refer to Figure 5E, the control circuit 130 is also used to calculate the total distribution interval of the first optimized distance ODa-1~ODa-n and the second optimized distance ODb-1~ODb-n on the time axis, and the first optimized distance ODa -1~ODa-n in the first distribution interval and second optimal distance on the time axis The second distribution interval from ODb-1~ODb-n on the time axis and the overlap interval between the first optimized distance ODa-1~ODa-n and the second optimized distance ODb-1~ODb-n on the time axis. Next, the control circuit 130 calculates the first width ratio between the first distribution interval and the total distribution interval, the second width ratio between the second distribution interval and the total distribution interval, and the third width ratio between the overlap interval and the total distribution interval.

In summary, the control circuit 130 can be used to determine the patient’s current behavior including: the number of zero-crossing points, the ratio of the area above and below the time axis of the curve, the sum of slopes, the first width ratio, and the second width ratio. And the third width ratio. The control circuit 130 can input at least one of the multiple characteristics to the classifier containing the care data 132 to compare with multiple sample points in the care data 132 to generate a comparison result. Then, the control circuit 130 judges whether the patient's current behavior belongs to one of multiple out-of-bed behaviors according to the comparison result. In some embodiments, the aforementioned classifier is a K-nearest neighbor algorithm classifier, but this disclosure is not limited to this. Classifiers that can compare the aforementioned multiple features and care data 132 are all within the scope of this disclosure.

In summary, the care system 100 in Figure 1 does not need to upload to the central server when judging the behavior of getting out of bed, and can perform classification, prediction, etc. in the device near the care recipient. Therefore, the care system 100 can avoid system failures and delays in judgment caused by network congestion, failure, or insufficient bandwidth, thereby ensuring that the caregiver can receive the alert for getting out of bed in real time.

FIG. 6 is a schematic diagram of a care system 600 according to an embodiment of the present disclosure. The care system 600 in Figure 6 is similar to the photo in Figure 1. The difference of the care system 100 is that the care system 600 replaces the position sensing system 120A with the position sensing system 120B. The position sensing system 120B is configured to be disposed on the plate-shaped member 105a or the plate-shaped member 105b, that is, the sensing range of the position sensing system 120B is aligned with the space 610 of the frame 101 for accommodating the patient. The position sensing system 120B is used to capture multiple images of the patient in the space 610 and provide the captured multiple images to the control circuit 130 as relative position information. The control circuit 130 performs calculations based on the relative position information and the care data 132 stored therein in advance to determine whether the patient's current behavior corresponds to one of the preset multiple out-of-bed behaviors.

FIG. 7 is a simplified functional block diagram of the position sensing system 120B according to an embodiment of the present disclosure. The position sensing system 120B includes one or more thermal imaging circuits 710 substantially aligned with the space 610, a microprocessor 720, and a communication interface 730, wherein the microprocessor 720 is coupled to the thermal imaging circuit 710 and the communication interface 730, and the communication interface 730 is used to communicate with the control circuit 130. In some embodiments, the thermal imaging circuit 710 may include one or more thermal imaging lenses.

In this embodiment, the control circuit 130 generates different behavior probabilities based on a single image and multiple images captured by the position sensing system 120B, and then comprehensively determines whether the patient has the behavior of getting out of bed based on these behavior probabilities. For example, the control circuit 130 executes the method 800A in FIG. 8A to generate the behavior probability corresponding to a single image. In the process S802A, the control circuit 130 selects one of the multiple images captured by the position sensing system 120B. The following uses the image 910 in FIG. 9A as a representative for description. In the process S804A, the control circuit 130 extracts the following features: image 910 The average of the global temperature and the median of the global temperature of multiple pixels.

In the process S806A, the control circuit 130 performs deback processing on the image 910. For example, the control circuit 130 can use multiple images to calculate the average temperature of the pixels at each position in a continuous period of time (hereinafter referred to as the first temperature threshold), that is, each pixel in the image 910 has an independent The first temperature threshold. The control circuit 130 regards the pixels in the image 910 that are lower than its first temperature threshold as the background and filters them out to obtain the foreground image 920 shown in FIG. 9B, in which the background marked in black represents the removed background. Then, in process S808A, the control circuit 130 extracts the following features: the average temperature of the foreground image 920, the median temperature of the foreground image 920, the number of pixels in each pixel row of the foreground image 920, and the temperature in the foreground image 920 The number of pixels above a second temperature threshold (that is, pixels marked with a white background). In some embodiments, the second temperature threshold is higher than the first temperature threshold of each pixel. Then, in the process S810A, the control circuit 130 inputs at least one of the multiple features extracted in the process S804A and the process S808A into the classifier to obtain the first behavior probability PBa corresponding to a single image.

In some embodiments, the control circuit 130 may input at least one of the multiple features extracted in the process S804A and the process S808A into the four classifiers in the process S810A to obtain the first candidate probability, the second candidate probability, and the first candidate probability, respectively. The probability of three candidates and the probability of fourth candidates. These four classifiers have been trained by machine learning for the four common directions of getting out of bed (for example, head left, head right, foot left, and right of the bed) in advance, so they respectively include the frame of the care data 132 corresponding to the patient Four directions of movement on 101 Sample points. Then, the control circuit 130 may select the first candidate probability, the second candidate probability, the third candidate probability, and the fourth candidate probability that has the largest value as the first behavior probability PBa.

The control circuit 130 also executes the method 800B in FIG. 8B to generate behavior probabilities corresponding to multiple images. The process S802B and the process S804B are similar to the process S802A and the process S806A respectively, and for the sake of brevity, the details are not repeated here. In the process S806B, the control circuit 130 subtracts the foreground image 920 from the multiple images captured by the position sensing system 120B, respectively, to obtain multiple temperature area changes corresponding to different time points. Then, in the process S808B, the control circuit 130 inputs a plurality of temperature area changes to the classifier to obtain the second behavior probability PBb-1 corresponding to the plurality of images.

In some embodiments, the control circuit 130 also executes the method 800C in FIG. 8C to generate behavior probabilities corresponding to multiple divided images. In the process S802C, the control circuit 130 selects one of the multiple images captured by the position sensing system 120B, such as the image 910. Then, in the process S804C, the control circuit 130 divides the image 910 into a left half and a right half. In the process S806C, the control circuit 130 will deback the left half and the right half of the image 910 in a method similar to the process S806A to generate the left half of the foreground image and the right half of the foreground image. For the sake of brevity, I won't go into details here.

In the process S808C, the control circuit 130 subtracts the left half of the foreground image from the left half of the multiple images captured by the position sensing system 120B, and subtracts the right half of the foreground image from the position sensing system 120B. The right halves of the captured images are subtracted separately to obtain multiple temperature area changes corresponding to different time points in the left and right regions of the image. Then, in the process S810C, the control circuit 130 inputs multiple temperature area changes corresponding to different time points in the left and right regions to the classifier to obtain the second behavior probabilities corresponding to the left half of the multiple images. PBb-2 and the second line probability PBb-3 corresponding to the right half of the multiple images. The method 800C can know the patient's body movement tendency (for example, to the left or right), and at the same time help to reduce the misjudgment caused by the shaking of the patient's hand.

In some embodiments, the control circuit 130 may compare multiple temperature area changes with a look-up table, multiple predetermined judgment rules, or multiple known sample points in process S808B or process S810C to obtain The second behavior probability is PBb-1, PBb-2, or PBb-3.

The control circuit 130 sums the first behavior probability PBa and the second behavior probability PBb-1, PBb-2, and PBb-3 to obtain a total result. Then the control circuit 130 will determine whether the total result exceeds a probability threshold, where the probability threshold can be stored in a memory circuit (not shown) of the control circuit 130 in advance. If the total result exceeds the probability threshold, the control circuit 130 will determine that the patient's current behavior corresponds to one of the preset multiple out-of-bed behaviors, and send a warning message to the main control device 12 and the cloud server in Figure 10 below. 16 and medical staff's mobile device 18. Conversely, the control circuit 130 can repeatedly execute the methods 800A, 800B, and 800C.

In some embodiments, the control circuit 130 may execute the method 800A and the method 800B, but the execution of the method 800C is omitted. In this way, The control circuit 130 only uses the first behavior probability PBa and the second behavior probability PBb-1 to determine the patient's current behavior. In other embodiments, the control circuit 130 may execute the method 800A and the method 800C, but the execution of the method 800B is omitted. In this way, the control circuit 130 only uses the first behavior probability PBa and the second behavior probability PBb-2 and PBb-3 to determine the patient's current behavior.

In summary, the control circuit 130 can subtract at least a part of the image 910 from at least a part of each of the multiple images captured by the position sensing system 120B to determine the patient's current behavior.

In still other embodiments, the method 800C may divide the image into multiple parts in the process S804C, so as to generate multiple second behavior probabilities corresponding to the multiple parts in the subsequent process, and the division direction does not need to be Special restrictions.

It is worth noting that the division of the image 910 into two in the method 800C is only an exemplary embodiment, and the number and direction of the image division can be adjusted according to actual analysis requirements. In practice, the classifier in the method 800A, 800B, or 800C can be a random forest classifier or a support vector machine.

In summary, the care system 600 in Figure 6 uses low-resolution thermal imaging to detect the patient’s movements, which can improve the accuracy of the identification of getting out of bed without infringing on the patient’s privacy, and can detect the patient’s getting out of bed earlier. Buy more time for frontline nursing staff to deal with out-of-bed incidents.

FIG. 10 is a simplified functional block diagram of the patient management system 10 according to an embodiment of the present disclosure. The patient management system 10 includes the main control device The device 12 is connected to a plurality of care systems (for example, the care systems 14-1 to 14-3), wherein the care systems 14-1 to 14-3 are connected to the main control device 12 through network communication. FIG. 11 is a schematic diagram of the host user interface 20 displayed by the host device 12. The host user interface 20 includes a plurality of display areas 22-1 to 22-3 respectively corresponding to the care systems 14-1 to 14-3. For the convenience of description, the following will use the care system 14 to refer to any unspecified one of the care systems 14-1 to 14-3, and use the display area 22 to refer to unspecified any of the display areas 22-1 to 22-3. One. The number of care systems in Figure 10 is only an example for ease of description, and the actual number of care systems can be adjusted according to the actual needs of the medical institution.

In some embodiments, the main control device 12 may be a personal computer, a notebook computer or a server installed in a nursing station of a medical institution. Each of the care systems 14-1 to 14-3 can be implemented by the care system 100 in FIG. 1 or the care system 600 in FIG. 6. That is, when the care system 14-1 to 14-3 determines that the patient's current behavior conforms to one of multiple preset bed-out behaviors, the care system 14-1 to 14-3 will send a warning message to the main control device 12.

As shown in Figure 11, the display area 22 is used to display the current behavior of the patient as judged by the care system 14. For example, the patient in bed 117 is putting down the handrail; the patient in bed 256 has moved to the end of the bed and has been out of bed. ; And the patient in bed 300 is getting up and so on. The display area 22 also includes a sensitivity adjustment pattern SP. The sensitivity adjustment pattern SP is a user-interactive object, and the main control device 12 is used to detect the cursor click corresponding to the position of the sensitivity adjustment pattern SP. When the cursor clicks and moves continuously, the sensitivity adjustment chart The case SP will move accordingly. In other words, the user of the main control device 12 can use the cursor to drag the sensitivity adjustment pattern SP.

The sensitivity adjustment pattern SP can move between three preset positions of low sensitivity, medium sensitivity, and high sensitivity. The three preset positions respectively represent different timings for the control circuit 130 to send warning messages to the main control device 12. For example, when the sensitivity adjustment pattern SP is at a high sensitivity position, the care system 14 will send a warning message when it determines that the patient's current behavior is "get up". For another example, when the sensitivity adjustment pattern SP is at the middle sensitivity position, the care system 14 will send a warning message when determining that the patient's current behavior is "moving to the end of the bed." For example, when the sensitivity adjustment pattern SP is at a low sensitivity position, the care system 14 will send a warning message when it determines that the patient's current behavior is "sitting on the edge of the bed."

In a word, the control circuit 130 will send a warning message when the patient’s current behavior is consistent with a kind of getting out of bed behavior, and by dragging the sensitivity adjustment pattern SP, the kind of getting out of bed behavior recorded by the control circuit 130 can be changed to another kind of getting out of bed behavior, in order to deal with different kinds of getting out of bed behavior. The patient of the care system 14 is adaptively set up in a situation where a bed-out warning is issued. When the main control device 12 receives the warning message, the "out of bed" icon in the display area 22 may display a predetermined color, flashing, or a combination of the two.

Please refer to FIG. 10 again. In some embodiments, the patient management system 10 further includes a cloud server 16 and a mobile device 18. The cloud server 16 and the mobile device 18 can be connected to the main control device 12 and the care systems 14-1 to 14-3 through network communication. The warning messages of the care system 14-1~14-3 will not only be sent to the main control device 12, but also input to the cloud server at the same time 16. The mobile device 18 is used to search for multiple times (for example, periodically) whether there is a new warning message input to the cloud server 16. If so, the mobile device 18 will obtain the relevant information of the new warning message, such as the actual geographic location of the care system 14 that sent the new warning message, the IP location, and the patient's medical history.

The mobile device 18 can send a connection request through the video and audio capture device 140 of the acquired IP location monitoring system 14. When the video and audio capture device 140 receives a request from the mobile device 18, the video and audio capture device 140 will turn on the camera, speaker and microphone to capture one or more images and audio of the patient, and the captured images The audio and video are transmitted to the mobile device 18 in the form of image stream and audio stream, so that the medical staff can understand the current situation of the patient through the mobile device 18 and make a remote call.

In practice, the mobile device 18 can be implemented by a smart phone, a tablet computer, or a head-mounted device (HMD for short). In some embodiments in which the mobile device 18 is implemented as a head-mounted device, as shown in Figure 12, the mobile device 18 may use Augmented Reality or Mixed reality technology. Within the user's field of view 30, the relevant information of the warning message is provided to the user through the mobile user interface 40. The mobile user interface 40 includes geographic location information 42 of the care system 14, medical record information 44 of patients in the care system 14, and a real-time image area 46. The real-time image area 46 is used to display one or more images received by the mobile device 18 from the audio-visual capture device 140 through the image stream.

In summary, the patient management system 10 provides diversified management for obtaining information. Roads and platforms enable medical staff to grasp the patient’s out-of-bed alarm in real time, whether they are on duty at the nursing station, performing daily patrols, or preparing medicines, thereby comprehensively improving the safety of the patient’s hospitalization.

In the specification and the scope of the patent application, certain words are used to refer to specific elements. However, those with ordinary knowledge in the technical field should understand that the same element may be called by different terms. The specification and the scope of patent application do not use the difference in names as a way of distinguishing components, but the difference in function of the components as the basis for distinguishing. The "including" mentioned in the specification and the scope of the patent application is an open term, so it should be interpreted as "including but not limited to". In addition, "coupling" here includes any direct and indirect connection means. Therefore, if it is described in the text that the first element is coupled to the second element, it means that the first element can be directly connected to the second element through electrical connection, wireless transmission, optical transmission, or other signal connection methods, or through other elements or connections. The means is indirectly connected to the second element electrically or signally.

The description method of "and/or" used herein includes any combination of one or more of the listed items. In addition, unless otherwise specified in the specification, any term in the singular case also includes the meaning of the plural case.

The above are only preferred embodiments of the present disclosure, and all equal changes and modifications made in accordance with the requirements of the present disclosure should fall within the scope of the disclosure.

100: care system

101: Frame

103a, 103b, 103c, 103d: armrest

105a, 105b: plate-shaped member

107: Edge

110: Cross-border detection system

111: The first distance sensor

112: The second distance sensor

120A: position sensing system

122: The first position sensor

124: second position sensor

130: control circuit

132: Care Information

140: Audio and video capture device

150: first sensing area

160: second sensing area

EV: Event notification

Da-1~DA-n: first distance

Db-1~Db-n: second distance

D1: First direction

D2: second direction

Claims (19)

A care system for predicting the behavior of getting out of bed, suitable for a frame used to support a patient, wherein an armrest is coupled to an edge of the frame parallel to a first direction, and a plate-shaped member is coupled to the frame Parallel to the other edge in a second direction, and the care system includes: at least one cross-border detection system, including: a first distance sensor for coupling to the armrest, and having a first sensor Area; and a second distance sensor having a second sensing area for coupling to the edge of the frame parallel to the first direction, or for coupling to the plate-shaped member and substantially Parallel and relative to the first distance sensor, wherein when the armrest is in an upright state, the first sensing area and the second sensing area at least partially overlap; a position sensing system for coupling In the frame, and used to obtain a relative position information between the patient and the frame; and a control circuit, used for data communication with the at least one cross-border detection system and the position sensing system, and used for Access a care data used to define the patient’s multiple out-of-bed behaviors on the frame; wherein when at least one of the following situations occurs, the control circuit sends a warning message: (1) The control circuit is based on the relative position The information and the care data determine that a current behavior of the patient corresponds to one of the multiple out-of-bed behaviors; (2) the at least one cross-border detection system senses that an object passes through. The care system according to claim 1, wherein the first The distance sensor includes a three-axis accelerator, and when the three-axis accelerator senses that the armrest changes from the upright state to a lowered state, the control circuit transmits the warning message. The care system according to claim 1, wherein the first distance sensor and the second distance sensor each include one or more infrared transceiving circuits. The care system according to claim 1, wherein the position sensing system comprises: a first position sensor for obtaining the relationship between the first position sensor and the patient substantially corresponding to the frame A plurality of first distances in the upper half of the frame; and a second position sensor for obtaining a plurality of first distances between the second position sensor and the patient substantially corresponding to the lower half of the frame Two distances; wherein the relative position information includes the plurality of first distances and the plurality of second distances. The care system according to claim 4, wherein the control circuit is configured to execute the following first optimization operation and a second optimization operation multiple times, so as to respectively generate a plurality of items on a scale axis in units of time. The first optimization distance and the multiple second optimization distances: the first optimization operation: select the consecutive Mth of the multiple first distances A distance, and the one with a predetermined characteristic among the consecutive M first distances is used as one of the plurality of first optimized distances, where M is a positive integer; the second optimization operation: select among the plurality of second distances The continuous M second distances, and the one with another predetermined characteristic in the continuous M second distances is used as one of the plurality of second optimization distances; wherein the control circuit is used to optimize according to the plurality of first optimization distances The distance, the plurality of second optimized distances, and the care data determine the current behavior of the patient. The care system according to claim 5, wherein the control circuit is configured to subtract the plurality of first optimized distances from the plurality of second optimized distances to obtain a plurality of coordinate points on a time axis, and the control circuit is configured to It is configured to calculate a plurality of characteristics according to the plurality of coordinate points, so as to determine the current behavior of the patient according to at least one of the plurality of characteristics and the care data. The care system according to claim 6, wherein the plurality of features include: the number of zero-crossing points of a curve formed by the plurality of coordinate points; surrounded by the curve and a time axis The ratio of the area of the graph above the time axis to the area of the graph below the time axis in a graph of The sum of the slopes of all connections. The care system according to claim 6, wherein the plurality of features include: a total distribution interval of the plurality of first optimized distances and the plurality of second optimized distances on the time axis; the plurality of first optimized distances The ratio between a first distribution interval on the time axis and the first width of the total distribution interval; the plurality of second optimized distances are between a second distribution interval on the time axis and the second width of the total distribution interval Ratio; and a third width ratio between an overlapping interval of the plurality of first optimized distances and the plurality of second optimized distances on the time axis and the total distribution interval. The care system according to claim 7 or 8, wherein the control circuit is configured to input the at least one of the plurality of characteristics into a classifier containing the care data, and the classifier calculates and compares the plurality of characteristics The at least one of and multiple sample points in the care data are used to generate a comparison result, and the current behavior of the patient is determined according to the comparison result. The care system according to claim 4, wherein the first position sensor and the second position sensor are ultrasonic sensors. The care system according to claim 1, wherein the position sensing system comprises: a thermal imaging circuit for capturing images corresponding to the frame for accommodating the disease Multiple images in a space of the patient, wherein the relative position information includes the multiple images. The care system according to claim 11, wherein the control circuit is configured to extract at least one of the following characteristics from a plurality of pixels in one of the plurality of images, so as to determine the characteristics of the plurality of images according to the At least one and the care data determine the current behavior of the patient: a global temperature average of the plurality of pixels; a global median temperature of the plurality of pixels; a temperature average of a foreground image, where each An average temperature of each pixel in a continuous period of time is set as a first temperature threshold of the pixel, and the temperature of each pixel of the foreground image is higher than the first temperature threshold; a median temperature of the foreground image Number; a total number of pixels in the foreground image; the number of pixels in each pixel column of the foreground image; and the number of pixels in the foreground image whose temperature is higher than a second temperature threshold, wherein the second temperature threshold is higher than each The first temperature threshold of the pixel. The care system according to claim 12, wherein the control circuit is configured to subtract at least a part of the foreground image and at least a part of each of the plurality of images respectively to obtain a plurality of temperatures corresponding to a plurality of time points The area change amount, and the control circuit is also configured to perform the following operations: generate a first step according to the at least one of the plurality of characteristics and the care data A behavior probability; generating one or more second behavior probabilities based on the plurality of temperature area changes and the care data; and if the sum of the first behavior probability and the one or more second behavior probabilities is greater than a probability threshold, It is determined that the current behavior of the patient corresponds to the one of the multiple out-of-bed behaviors. The care system according to claim 13, wherein the control circuit is configured to input the at least one of the plurality of characteristics to one or more classifiers to obtain the first behavior probability, and the control circuit is further configured To input the multiple temperature area changes into another classifier or compare with a look-up table, multiple pre-made judgment rules or multiple known sample points to obtain the one or more second behaviors Probability. The care system according to claim 14, wherein the one or more classifiers include four classifiers, and the four classifiers respectively include four moving directions on the frame corresponding to the patient in the care data The control circuit is set to select the largest among a first candidate probability, a second candidate probability, a third candidate probability, and a fourth candidate probability outputted by the four classifiers as the first Probability of behavior. A patient management system including: One or more care systems, each of which is suitable for a frame for supporting a patient, and includes: at least one cross-border detection system for coupling to the frame; and a position sensing system for Coupled to the frame and used to obtain a relative position information between the patient and the frame; and a control circuit for data communication with the at least one cross-border detection system and the position sensing system, and It is used to access a care data used to define the patient’s multiple out-of-bed behaviors on the frame, wherein when at least one of the following situations occurs, the control circuit sends a warning message: (1) The control circuit is based on the The relative position information and the care data determine that a current behavior of the patient corresponds to one of the multiple out-of-bed behaviors; (2) the at least one out-of-bounds detection system senses the passage of an object; and a main control device with It communicates with the control circuit to receive and correspondingly display the warning message, and is used to provide a host user interface, wherein the host user interface includes one corresponding to one or more care systems. Or a plurality of sensitivity adjustment patterns; wherein each sensitivity adjustment pattern is a user-interactive object, when the sensitivity adjustment pattern is moved from a first preset area to a second preset area, the one or more caregivers The control circuit corresponding to the sensitivity adjustment pattern in the system can change the one of the multiple exiting behaviors to the other of the multiple exiting behaviors. The patient management system as described in claim 16, further comprising: A cloud server, wherein the cloud server is used to receive and store the warning message from the control circuit; a mobile device is used to access the cloud server to determine whether the cloud server stores the warning message; and a Video capture device for wireless communication with the mobile device; wherein when the mobile device determines that the cloud server receives the warning message, the mobile device is set to perform the following operations: instruct the one or more care systems The audio-visual capture device in the care system corresponding to the warning message captures one or more images, and transmits the one or more images to the mobile device as an image stream; and the care system corresponding to the warning message The audio-visual capture device of the system performs voice communication. The patient management system according to claim 16, wherein an armrest is coupled to an edge of the frame in the first direction, and a plate-shaped member is coupled to the other edge of the frame in the second direction, and the At least one cross-border detection system includes: a first distance sensor for coupling to the armrest and having a first sensing area; and a second distance sensor for coupling to the frame The edge in one direction may be used for coupling to the plate-shaped member and has a second sensing area; wherein when the armrest is in a standing state, the first sensing area and the second sensing area The areas overlap at least partially. The patient management system according to claim 17, wherein the mobile device is used to display a mobile user interface, and the mobile user interface includes: a geographic location of the care system corresponding to the warning message; A medical record information of a patient in the care system of the warning message; and a real-time image area for displaying the one or more images received by the mobile device through the image stream.

Images