TWI771838B - Wound multiple sensing method and wound multiple sensing system - Google Patents

Abstract

The present invention discloses a wound multiple sensing method, including: calculating a similarity between the current data sequence and each of the case data sequences for each of the reference cases; selecting the case data sequence which has the highest similarity with the current data sequence, from the case data sequences for each of the reference cases, to be a similar case data sequence in each of the reference cases, wherein the similar case data sequence is corresponding to a similar case treatment; performing a multiple regression analysis using the similar case data sequences and the similar case treatments to calculate a fitness function, wherein the dependent variable of the fitness function is a wound change; performing a parameter optimization algorithm using the current data sequence and the fitness function to calculate a best treatment maximizing the wound change, and an expected wound change value corresponding to the best treatment.

Description

Wound multiple sensing method and system

The present invention relates to a sensing method and system, in particular to a wound multiple sensing method and system.

According to statistics, the global prevalence of arteriovenous ulcers and pressure ulcers in the elderly population is 6%, and the global prevalence of diabetes is 8.8%. These patients often develop chronic wounds that are difficult to heal for more than three months, showing a great need for remote wound care.

General nurses or nurses who perform wound care usually only make judgments and treatments based on the current condition of the wound, such as the most basic wound cleaning and disinfection. However, this kind of simple nursing and treatment of "headache, foot pain, foot pain" often makes the wound stay in the inflammatory or proliferative stage for a long time, delaying the recovery of the wound. What's more, wounds fail to heal due to misjudgment, increasing the risk of infection or necrosis, which can ultimately result in sepsis or amputation.

In addition, due to the complex and multi-faceted nature of care recommendations (guidelines), it is difficult for general nurses or caregivers to make extensive judgments. In the face of complex situations, dogmatic guidance often leads to unclear or even conflicting situations, which makes it even more difficult for nurses or caregivers to deal with patient wounds. In addition, due to the wide variety of dressing products, there are often difficulties in selection, so the experience of professionals, such as Certified Wound, Ostomy and Continence Nurse (CWOCN), needs to be adopted, and more knowledge of wound conditions is required. Observations and recommendations.

Therefore, there is a need for a wound multi-sensing method and system that can estimate the progress of wound healing and provide recommendations on care and treatment.

An embodiment of the present invention provides a wound multi-sensing method, comprising: reading a plurality of reference cases, wherein each reference case includes a plurality of case data sequences; receiving a current wound representation of the wound, wherein the current wound representation includes a plurality of wound observation data ; Based on multiple wound observation data of the wound, establish the current data sequence; Calculate the similarity between the current data sequence and each case data sequence of each reference case; The case data sequence with the highest similarity is used as the similar case data sequence in each reference case, where the similar case data sequence corresponds to the similar case treatment policy; multiple regression analysis is performed using the similar case data sequence and the similar case treatment policy ), to calculate the fitness function, wherein the strain number of the fitness function is the wound change; use the current data sequence and the fitness function to perform a parameter optimization algorithm to calculate the maximum wound change that maximizes The optimal treatment policy and the expected wound change value corresponding to the optimal treatment policy.

In some embodiments, the above method further includes: first operation: using the current data series and the expected wound change value to estimate the expected data series, wherein the expected data series includes at least one size parameter; the second operation: using the expected data series and adapting The degree function performs a parameter optimization algorithm to calculate the next best treatment policy that maximizes the wound change and the next expected wound change value corresponding to the next best treatment policy, and then uses the expected data sequence and the next expected wound The change value is used to estimate the next expected data sequence; wherein, the next expected data sequence is used by the processor as a new expected data sequence to repeat the second operation until it is calculated that the at least one size parameter in the next expected data sequence is all 0.

In certain embodiments, the at least one dimension parameter is expected wound length, expected wound width, and expected wound depth.

In some embodiments, the above method further includes: accumulating the number of times the second operation is performed until the at least one size parameter in the next expected data sequence is calculated to be 0; adding 1 to the number of times the second operation is performed , multiplied by the sensed period to get the expected wound healing days.

In certain embodiments, the wound change comprises the following three dimensions: wound length change, wound width change, and wound depth change; and wherein maximizing wound change comprises: maximizing wound volume change; and wherein the expected wound change value comprises Expected Wound Length Change, Expected Wound Width Change, and Expected Wound Depth Change.

In some embodiments, the above-mentioned method further comprises: using a color sensor in the wound machine to obtain wound tissue data among the plurality of wound observation data; using a depth sensor in the wound machine to obtain a plurality of wound observation data Wound area surface area data; use the temperature sensor in the wound machine to obtain wound temperature data from multiple wound observations.

In some embodiments, each reference case further includes a characterization record, a qualitative questionnaire record, and a care disposition record; wherein the case data sequence is established based on the characterization record; and wherein the similar case disposition policy is based on the qualitative questionnaire record and care treatment record Disposition records are established.

Embodiments of the present invention provide a wound multi-sensing system, the system including a processor for performing: reading a plurality of reference cases, wherein each reference case includes a plurality of case data sequences; receiving a current wound representation of the wound, wherein The current wound representation includes multiple wound observation data; based on the multiple wound observation data of the wound, the current data sequence is established; the similarity between the current data sequence and each case data sequence of each reference case is calculated; from the case data sequence of each reference case Select the case data sequence with the highest similarity with the current data sequence as the similar case data sequence in each reference case, where the similar case data sequence corresponds to the similar case handling policy; use the similar case data sequence and the similar case handling policy Perform a multiple regression analysis to calculate a fitness function, where the strain number of the fitness function is the wound change; use the current data sequence and the fitness function to perform a parameter optimization algorithm to calculate the optimal wound to maximize the change. The treatment policy and the expected wound change value corresponding to the optimal treatment policy.

In some embodiments, the processor further performs: a first operation: using the current data series and the expected wound change value to estimate an expected data series, wherein the expected data series includes at least one size parameter; the second operation: using the expected data series and The fitness function performs a parameter optimization algorithm to calculate the next best treatment policy that maximizes the wound change and the next expected wound change value corresponding to the next best treatment policy, and then uses the expected data sequence and the next expected The wound change value is used to estimate the next expected data sequence; wherein, the next expected data sequence is used as a new expected data sequence to repeat the second operation until the at least one size parameter in the next expected data sequence is calculated to be 0.

In some embodiments, the processor further executes: accumulating the number of times the second operation is performed until it is calculated that the at least one size parameter in the next expected data sequence is all 0; the number of times the second operation is performed Add 1 and multiply by the period sensed to get the expected wound healing days.

In some embodiments, the system further includes a wound machine, and the wound machine includes: a color sensor for acquiring wound tissue data in a plurality of wound observation data; a depth sensor for acquiring a plurality of wound observation data in Wound area surface area data; temperature sensor, used to obtain wound temperature data in multiple wound observation data.

In some embodiments, the plurality of wound observation data includes wound tissue data, wound area surface area data and wound temperature data, and the system further includes a wound machine, the wound machine including: a color sensor for obtaining wound tissue data; depth The sensor is used to obtain the surface area data of the wound area; the temperature sensor is used to obtain the wound temperature data.

The wound multi-sensing method and wound multi-sensing system provided by the present invention can estimate the progress of wound healing based on the current condition of the wound by referring to similar cases accumulated in the past, and provide suggestions on care and treatment.

The wound multi-sensing method provided by the present invention compares the current condition of the patient's wound with a plurality of reference cases to screen out a plurality of similar cases, and then based on the treatment policy of the similar cases and the course of wound healing progress, estimate the progress of wound healing, and provide advice on care and management.

According to an embodiment of the present invention, each of the above reference cases may include a plurality of evaluation results, the number of which depends on the time interval from when the first wound is sensed to when the wound is fully healed, and The period during which the wound is sensed. For example, take the first day when the wound is sensed as the first day, and the wound is completely healed on the 18th day. If the wound is sensed once a day during the 18 days, the reference case will have 18 Pen assessment results; if the wound is sensed once every three days within the 18 days, the reference case will have 6 assessment results. Each evaluation result can also include three parts: characteristic records, qualitative questionnaire records, and nursing treatment records.

The above characterization records may include, for example, wound location (e.g. back, ankle, wrist, etc.), wound length, wound width, wound depth, wound extent surface area, amount of exudate, nature of exudate, color of exudate, wound temperature, wound tissue ...and many other wound observation data. In some embodiments, a wound machine may be used to obtain multiple wound observations in a characterization record of a patient's wound. For example, the color sensor in the wound machine can be used to obtain wound tissue data, the depth sensor in the wound machine can be used to obtain wound area surface area data, and the temperature sensor in the wound machine can be used to obtain wound temperature data .

The above-mentioned qualitative questionnaire records may include, for example, the bed resting environment (such as the frequency and method of making sheets and bedding), turning techniques (such as the frequency and method of turning over), wound care (such as the frequency and method of changing dressings), and physiological observations (such as wound Visual assessment of the condition, presence or absence of symptoms of fever or chills, and assessment of other physiological conditions), etc. The above-mentioned treatment records may include, for example, the type of dressing (such as gauze, silicone, foam, etc.), the object of health education (such as nursing or family members), and the content of health education (such as bedding and bedding techniques, wound care techniques, turning over techniques, etc. )Wait.

In some embodiments, the qualitative questionnaire record and treatment record are filled out by the caregiver or the patient based on their actual experience, cognition and feeling of the care process.

FIG. 1 is a flowchart of a method 100 according to an embodiment of the present invention. As shown in FIG. 1, method 100 includes operations 101-107. In operation 101, a plurality of reference cases are read, wherein each reference case includes a plurality of case data sequences. The method 100 then proceeds to operation 102 .

According to an embodiment of the present invention, as described above, the number of reference cases read in operation 101 depends on the time interval from when the wound is sensed for the first time to when the wound is fully healed, and when The period during which the wound is sensed. The multiple case data sequences included in each reference case are established based on multiple wound observation data recorded for each sensed characterization.

FIG. 2 is an example of a reference case 200 provided according to an embodiment of the present invention. As shown in FIG. 2 , the reference case 200 includes a plurality of case data sequences, such as a case data sequence 201 , a case data sequence 202 , a case data sequence 203 , and a case data sequence 204 . In the example of the present disclosure, it is assumed that the time from when the wound is sensed for the first time (ie, "Day 1" in Figure 2) to the time when the wound is completely healed (ie, "Day N" in Figure 2) ), the wound was sensed every three days. Case data sequence 201, case data sequence 202, case data sequence 203, and case data sequence 204 are based on multiple wound observation data recorded on the first, fourth, seventh, and Nth days, respectively. Establish. For example, {2.0, 3.0, 0.5, 50, 32, . 50. The wound temperature is 32... etc.; {3.0, 2.0, 0.4, 25, 31, ...} of the case data sequence 202 can represent that the wound length sensed on the 4th day is 3.0, the wound width is 2.0, and the wound depth is 0.4, the amount of exudate is 25, the temperature of the wound is 31, etc.; {2.5, 1.5, 0.3, 30, 31, ...} of the case data sequence 203 can represent that the wound length sensed on the 7th day is 2.5, and the wound width is 1.5. The depth of the wound is 0.3, the amount of exudate is 30, the temperature of the wound is 31, etc.; all data in the case data sequence 201 is 0, which means that the wound is detected to be completely healed on the Nth day.

Returning to FIG. 1 , the method 100 continues with operation 102 . At operation 102, a current wound representation of the wound is received, wherein the current wound representation includes a plurality of wound observation data. The method 100 then proceeds to operation 103 .

According to an embodiment of the present invention, the multiple items of wound observation data included in the above-mentioned current wound characterization may correspond to the multiple items of wound observation data included in the characterization record of the aforementioned reference case. That is, the current wound characterization may include, for example, wound location (eg, back, ankle, wrist, etc.), wound length, wound width, wound depth, wound extent surface area, exudate volume, exudate properties, exudate color, wound temperature , wound tissue... and other wound observation data. Likewise, in some embodiments, a wound machine may be used to obtain multiple wound observations of the patient's wound in the current wound characterization. For example, the color sensor in the wound machine can be used to obtain wound tissue data, the depth sensor in the wound machine can be used to obtain wound area surface area data, and the temperature sensor in the wound machine can be used to obtain wound temperature data .

Returning to FIG. 1 , the method 100 continues with operation 103 . In operation 103, a current data sequence is established based on the multiple wound observation data of the wound. The method 100 then proceeds to operation 104 .

According to the embodiment of the present invention, the recording format of the above-mentioned current data sequence is also the same as that of the case data sequence of the aforementioned reference case. That is to say, the current data sequence can be recorded in the form of {x ₁ , x ₂ , x ₃ , x ₄ , x ₅ , ...}, like the case data sequence 201 , the case data sequence 202 or the case data in Figure 2 sequence 203.

In some embodiments, in operation 103, a plurality of wound observation data of the wound are first subjected to a standardization process, and then a current data sequence is created based on the standardized wound observation data. Since normalization is a well-known data processing technique, it will not be repeated here. In some embodiments, the range of each wound observation data can be converted to a range of a specific value range (eg, 0-5) through the operation of certain mathematical functions.

Returning to FIG. 1 , the method 100 continues with operation 104 . In operation 104, the similarity between the current data sequence and each case data sequence of each reference case is calculated. For example, calculate the current data series and the reference case 200 in Figure 2, the case data series 201 on the 1st day, the case data series 202 on the 4th day, the case data series 203 on the 7th day... (and so on) and calculate the similarity between the current data sequence and each case data sequence of each other reference case. The method 100 then proceeds to operation 105 .

。應注意的係，計算得出的歐氏距離越小，代表相似度越高。 In some embodiments, the above calculation of similarity is to regard the current data sequence and each case data sequence as two points on a coordinate system, and then calculate the Euclidean distance between the two points ( Euclidean Distance). Specifically, the current data series is subtracted from each value in a case data series to obtain multiple differences, then the sum of squares of these differences is calculated, and finally the square is opened to obtain the current data. The Euclidean distance between the sequence and the sequence of the case data. For example, if the current data sequence is {x ₁ , x ₂ , x ₃ , …, x _n }, and an example case data sequence is {y ₁ , y ₂ , y ₃ , …, y _n }, then the The Euclidean distance between the current data sequence and the case data sequence of this example is

. It should be noted that the smaller the calculated Euclidean distance, the higher the similarity.

Returning to FIG. 1 , the method 100 continues with operation 105 . In operation 105, from the case data sequence of each reference case, the case data sequence with the highest similarity with the current data sequence (eg, the smallest Euclidean distance) is selected as the similar case data sequence in each reference case. For example, from all the case data sequences of the reference case 200 in Figure 2, select the case data sequence with the highest similarity to the current data sequence (for example, the smallest Euclidean distance), as the similar case data in the reference case 200. sequence, and perform the same operation on other reference cases to obtain multiple similar case data sequences. The method 100 then proceeds to operation 106 .

According to an embodiment of the present invention, each of the above-mentioned similar case data sequences corresponds to a similar case handling policy. Similar case treatment policies are established based on qualitative questionnaire records and treatment records corresponding to the representation records used to establish similar case data sequences. For example, it is assumed that the similar case data sequence in the reference case 200 in Figure 2 is the case data sequence 203, which means that in the multiple evaluation results of the reference case 200, the wound representation on the 7th day is the same as the patient's current condition. The wounds of the patients were most similar, so the treatment strategy for similar cases was established based on the qualitative questionnaire records and treatment records on the 7th day.

Returning to FIG. 1 , the method 100 continues with operation 106 . In operation 106, multiple regression analysis is performed using the similar case data sequences and their corresponding similar case treatment policies to calculate a fitness function, wherein the strain factor of the fitness function is the wound change . The method 100 then proceeds to operation 107 .

FIG. 3A is an example of input data 300A required for multiple regression analysis according to an embodiment of the present invention. As shown in FIG. 3, input data 300A includes similar case data sequences 301, similar case treatment policies 302, and wound changes 303 from multiple reference cases. Wound changes 303 are obtained by comparing the differences between the similar case data sequence in each reference case and its previous case data sequence. For example, it is assumed that the similar case data sequence in the reference case 200 in Figure 2 is the case data sequence 203, and the row 304 in Figure 3A is the similar case data sequence in the reference case 200 (ie the case data sequence 203) , similar case treatment policy (that is, the case treatment policy established based on the qualitative questionnaire records and treatment records on the 7th day) and the combination of wound changes, the wound change value of row 304 is calculated by comparing the case data series 203 with the previous one. Differences between case data sequences (ie case data sequence 202) are obtained. In various embodiments, the wound change may be, for example, a change in wound length, a change in wound width, a change in wound surface area, or a change in wound volume, and the like. In the example shown in Figure 3A, wound change refers to wound volume change. For example, the value of the wound change in row 304 is the difference between the wound volume indicated by case data sequence 203 (ie, the product of wound length, wound width, and wound depth) and the wound volume indicated by case data sequence 202, such as The following calculation: 3.0*2.0*0.4 - 2.5*1.5*0.3 = 2.275

Returning to FIG. 1 , the method 100 continues with operation 107 . At operation 107, a parameter optimization algorithm is performed using the current data sequence and the fitness function to calculate an optimal treatment policy that maximizes the wound change and an expected wound change value corresponding to the optimal treatment policy. Optimal treatment guidelines and expected wound change values can be displayed on the display device to provide caregivers with advice on wound care management.

According to the embodiment of the present invention, the parameter optimization algorithm takes the current data sequence and the fitness function as input, and its function is to find the best case treatment policy, that is, the best qualitative questionnaire record data and treatment records The combination of data (eg, the combination of turning frequency, dressing, wound care technique, etc. in Figure 3A) maximizes wound changes (eg, changes in wound length, changes in wound width, changes in wound surface area, or changes in wound volume). The case treatment policy at this time is the best case treatment policy, and the value of the maximized wound change is the expected wound change value. For example, suppose there is a first treatment policy of "turning frequency 3 times, dressing with foam, excellent wound care technology..." and the corresponding wound change value is 3.5. If the first treatment policy is changed so that the wound change value is greater than 3.5, the first treatment policy is the optimal treatment policy, and the expected wound change value corresponding to the first treatment policy is 3.5. In some embodiments, the parameter optimization algorithm may be, for example, Genetic Algorithm (GA), Ant Colony Optimization (Ant Colony Optimization), Particle Swarm Optimization (PSO) and other heuristic algorithms, the present invention is not limited to this.

FIG. 3B is an example of input data 300B required for multiple regression analysis according to a preferred embodiment of the present invention. The difference between Fig. 3B and Fig. 3A is that the wound change 303 further includes three dimensions: wound length change, wound width change, and wound depth change, which means that the fitness function calculated by performing multiple regression analysis has wound length change , wound width change and wound depth change. In the example provided in this disclosure, the value of the wound length change (0.5) in row 304 is the difference between the wound length (2.5) indicated by the case data sequence 203 and the wound length (3.0) indicated by the case data sequence 202, Calculate as follows: 3.0 - 2.5 = 0.5 The value of the change in wound width (0.5) for row 304 is the difference between the wound width (1.5) indicated by case data series 203 and the wound width (2.0) indicated by case data series 202, calculated as follows: 2.0 - 1.5 = 0.5 The value of the change in wound depth (0.1) for row 304 is the difference between the wound depth (0.3) indicated by case data series 203 and the wound depth (0.4) indicated by case data series 202, calculated as follows: 0.4 - 0.3 = 0.1

In the preferred embodiment shown in Fig. 3B, since the fitness function has three strain numbers, such as wound length change, wound width change and wound depth change, the expected wound change value output by the parameter optimization algorithm is expected The combination of the wound length change value, the expected wound width change value and the expected wound depth change value, and the object maximized by the parameter optimization algorithm is the wound volume change. For example, assuming that the current wound length indicated by the current data sequence is 10, the wound width is 5, and the wound depth is 0.5, and there is a second treatment policy "turning frequency 3 times, the dressing is foam, and the wound care technique is excellent..." The corresponding change in wound length is 2, the change in wound width is 1, and the change in wound depth is 0.1, then the change in wound volume corresponding to the second treatment policy is: current wound length, current wound width and current wound The product of the three depths, minus the product of "the current wound length minus the change in the wound length", "the current wound width minus the change in the wound width", and "the current wound depth minus the change in the wound depth". , calculated as follows: 10*5*0.5 - (10-2)*(5-1)*(0.5-0.1) = 12.2 The parameter optimization algorithm determines that the change in wound volume cannot be changed by changing the second treatment policy to a value greater than 12.2, then the second treatment policy is the optimal treatment policy, and the expected change in wound length corresponding to the second treatment policy The value is 2, the expected wound width change value is 1, and the expected wound width is 0.1.

In some embodiments, after the operation 107 in the method 100 is performed, the method 400 is further performed. FIG. 4 is a flowchart of a method 400 according to an embodiment of the present invention, and FIG. 5 is a series of optimal treatment policies generated during the process of executing the method 400 provided by a preferred embodiment of the present invention , an example of a series of expected wound change values and a series of expected data series. The following description will need to refer to FIG. 4 and FIG. 5 in conjunction with each other for a better understanding of the present disclosure. It should be noted that the numerical values in each table in FIG. 5 are only used as examples to facilitate the description of the embodiments of the present invention, and are not intended to limit the present invention.

As shown in FIG. 4 , method 400 includes operations 401 - 404 , and method 400 begins by performing operation 107 and then proceeds to operation 401 . In operation 401, an expected data series is estimated using the current data series and the expected wound change value, wherein the expected data series includes at least one size parameter. Method 400 then proceeds to operation 402 .

According to an embodiment of the present invention, the above-mentioned expected data sequence is the result of subtracting the current data sequence and the expected wound change value. In some embodiments, the at least one size parameter included in the expected data sequence may be expected wound length, expected wound width, expected wound depth, or any combination of the above.

In a preferred embodiment, the fitness function has three strain numbers such as wound length change, wound width change and wound depth change, so the expected wound change value output by the parameter optimization algorithm includes the expected wound length change value, the expected wound length change value, Wound width change value and expected wound depth change value, and the size parameters included in the expected data series are expected wound length, expected wound width and expected wound depth. Taking Fig. 5 as an example, the current data sequence 501 established according to the wound condition at the current time point indicates that the current wound length is 10, the wound width is 5, and the wound depth is 0.5. After the operation 107 is performed, the optimal treatment policy 502 and the expected wound change value 503 are obtained, wherein the expected wound change value 503 includes the expected wound length change value (0.6), the expected wound width change value (0.3) and the expected wound depth change value (0.1). Then, through operation 401, the wound length (10), wound width (5) and wound depth (0.5) indicated by the current data sequence 501 are compared with the expected wound length change value (0.6), the expected wound length change value (0.6) in the expected wound change value 503, The wound width change value (0.3) and the expected wound depth change value (0.1) are subtracted to obtain the expected data sequence 504, which is calculated as follows: Expected wound length for expected data sequence 504: 10 - 0.6 = 9.4 Expected Wound Width for Expected Data Sequence 504: 5 - 0.3 = 4.7 Expected wound depth for expected data series 504: 0.5 - 0.1 = 0.4

Returning to FIG. 4 , method 400 continues with operation 402 . In operation 402, a parameter optimization algorithm is performed using the expected data sequence and the fitness function to calculate the next best treatment policy that maximizes the wound change and the next expected wound change value corresponding to the next best treatment policy. The method 400 then proceeds to operation 403 .

According to an embodiment of the present invention, operation 402 is similar to the parameter optimization algorithm and fitness function used in operation 107, except that operation 107 searches for the best treatment policy based on the current data sequence, and operation 402 is based on the expected data sequence to search for the next best solution. Taking Fig. 5 as an example, in this preferred embodiment, after the expected data sequence 504 has completed the operation 402, the next best treatment policy, namely the best treatment policy 505, and the next expected wound change value, namely Expected wound change value 506, wherein expected wound change value 506 includes expected wound length change value (0.8), expected wound width change value (0.5), and expected wound depth change value (0.1).

Returning to FIG. 4 , method 400 continues with operation 403 . At operation 403, the next expected data series is estimated using the expected data series and the next expected wound change value. Method 400 then proceeds to operation 404 .

According to an embodiment of the present invention, operation 403 is similar to operation 401 for estimating the expected data series, except that operation 401 subtracts the current data series from the expected wound change value, and operation 403 compares the expected data series to the next expected wound The change value is subtracted. Taking FIG. 5 as an example, in this preferred embodiment, the operation 403 is to subtract the expected data sequence 504 and the next expected wound change value obtained after the operation 402 is performed, that is, the expected wound change value 506, to perform a subtraction, The next expected data sequence, that is, the expected data sequence 507, is obtained, as calculated as follows: Expected wound length for expected data series 507: 9.4 - 0.8 = 8.6 Expected Wound Width for Expected Data Sequence 507: 4.7 - 0.5 = 4.2 Expected wound depth for expected data series 507: 0.4 - 0.1 = 0.3

Returning to FIG. 4 , method 400 continues with operation 404 . In operation 404, it is determined whether the at least one size parameter in the next expected data sequence obtained in the previous operation is all zero. If not, method 400 returns to operation 402 . If so, it means that wound healing is expected and method 400 ends.

Taking FIG. 5 as an example, in this preferred embodiment, the expected data sequence 507 will again perform operations 402 , 403 and 404 to obtain the next expected data sequence. Operation 402, operation 403 and operation 404 are repeatedly performed in this way until wound healing is expected. As shown in the expected data sequence 500 in FIG. 5, the expected wound length, expected wound width and expected wound depth are all zero.

According to an embodiment of the present invention, a series of optimal treatment policies (eg optimal treatment policy 502, optimal treatment policy 505, ... in FIG. 5), a series of expected wound change values ( For example, expected wound change value 503, expected wound change value 506...) in Figure 5, and a series of expected data series (eg, expected data series 504, expected data series 507...) in Figure 5, can be displayed on the display device to provide advice on wound care management. In some embodiments, the total number of times operations 402 and 403 are performed during the execution of the method 400 may be accumulated, and then the number of times operations 402 and 403 are performed is incremented by 1, and then multiplied by the period during which the wound is sensed , the expected number of days it takes for the wound to heal. For example, if the wound is sensed once every three days, and it is expected that operation 402 and operation 403 will be performed 8 times before the wound will heal, it can be expected that the number of days required for the wound to heal is 27 days, as calculated as follows: : (8+1)*3 = 27

FIG. 6A is a block diagram of a computing environment in which the system 600A is located according to an embodiment of the present invention. As shown in FIG. 6A, the system 600A includes a processor 601. The processor 601 can be connected to the storage device 602 in a wired or wireless manner to read reference cases stored in the storage device 602 to execute the aforementioned method. 100 and method 400. The processor 601 can also be connected to the display device 603 in a wired or wireless manner, so as to display a series of optimal treatment guidelines, a series of expected wound change values and a series of expected data series generated by executing the method 100 and the method 400 on the display device 603 to provide caregivers with advice on wound care management. It should be noted that although the storage device 602 in FIG. 6A is not included in the system 600A, in some embodiments, the system 600A may further include the storage device 602 . Similarly, although the display device 603 in FIG. 6A is not included in the system 600A, in some embodiments, the system 600A may further include the display device 603 .

The processor 601 can be any device for executing instructions, such as a central processing unit (CPU), a microprocessor (microprocessor), a controller, a microcontroller (microcontroller) or a state machine (state machine), and the present invention does not Limited to this; the storage device 602 can be any device for storing data, such as disk drives, optical storage devices, and solid-state storage devices such as random access memory (RAM), read only memory (ROM), The present invention is not limited to removing the media device, memory card or flash memory card; the display device 603 can be any device for displaying text and images, such as LCD display, LED display, OLED display or plasma Display, the present invention is not limited to this.

FIG. 6B is a block diagram of a computing environment in which the system 600B is located according to another embodiment of the present invention. Compared with the system 600A in FIG. 6A , the system 600B further includes a wound machine 604 , and the wound machine 604 further includes a color sensor 605 , a depth sensor 606 and a temperature sensor 607 for obtaining wound tissue data respectively. , wound area surface area data and wound temperature data. The wound machine 604 can be connected to the processor 601 in a wired or wireless manner, so as to transmit a plurality of obtained wound observation data to the processor 601, and then the processor 601 can establish a current data sequence based on the wound observation data. It should be noted that although the storage device 602 in FIG. 6B is not included in the system 600B, in some embodiments, the system 600B may further include the storage device 602 . Similarly, although the display device 603 in FIG. 6B is not included in the system 600B, in some embodiments, the system 600B may further include the display device 603 . In addition, although the system 600B includes the processor 601 and the processor 604, the present invention does not limit whether the processor 601 and the processor 604 are disposed in the same electronic device.

In conclusion, the wound multi-sensing method and wound multi-sensing system provided by the present invention can estimate the progress of wound healing based on the current condition of the wound by referring to similar cases accumulated in the past, and provide suggestions on care and treatment.

The serial numbers in this specification and the scope of the patent application, such as "first", "second", etc., are only for convenience of description, and there is no sequential relationship between them.

The above paragraphs use multiple levels of description. Obviously, the teachings herein can be implemented in a variety of ways, and any particular architecture or functionality disclosed in the examples is merely a representative case. Based on the teachings herein, anyone skilled in the art should understand that each aspect disclosed herein may be implemented independently or two or more aspects may be implemented in combination.

Although the present disclosure has been disclosed above with examples, it is not intended to limit the present disclosure. Anyone who is familiar with the art can make some changes and modifications without departing from the spirit and scope of the present disclosure. Therefore, the protection scope of the invention is The scope of the patent application attached herewith shall prevail.

100: Method 101-107: Operations 200: Reference Case 201-204: Case data series 300A: Input data 301: Similar case data sequence 302: Similar Case Handling Policy 303: Wound changes 304: OK 300B: Input information 400: Method 401-404: Operations 500,504,507: Expected data series 501: Current data sequence 502, 505: Best Practices 503,506: Expected Wound Variation Values 600A: System 601: Processor 602: Storage Device 603: Display device 600B: System 604: Wound Machine 605: Color Sensor 606: Depth Sensor 607: Temperature sensor

The present disclosure will be better understood from the following description of exemplary embodiments in conjunction with the accompanying drawings. In addition, it should be understood that, in the flowcharts of the present disclosure, the order of execution of various blocks may be changed, and/or certain blocks may be changed, omitted, or combined. FIG. 1 is a flowchart of a method 100 according to an embodiment of the present invention. FIG. 2 is an example of a reference case 200 provided according to an embodiment of the present invention. FIG. 3A is an example of input data 300A required for multiple regression analysis according to an embodiment of the present invention. FIG. 3B is an example of input data 300B required for multiple regression analysis according to a preferred embodiment of the present invention. FIG. 4 is a flowchart of a method 400 according to an embodiment of the present invention. FIG. 5 is an example of a series of optimal treatment guidelines, a series of expected wound change values, and a series of expected data series generated during the execution of method 400 according to a preferred embodiment of the present invention. FIG. 6A is a block diagram of a computing environment in which the system 600A is located according to an embodiment of the present invention. FIG. 6B is a block diagram of a computing environment in which the system 600B is located according to another embodiment of the present invention.

100: Method 101-107: Operations

Claims (14)

A wound multiple sensing method, executed by a processor, comprising: reading a plurality of reference cases, wherein each of the reference cases includes a plurality of case data sequences corresponding to different time points; receiving a current wound of a wound Characterization, wherein the current wound representation includes a plurality of wound observation data; based on the plurality of wound observation data of the wound, a current data sequence is established; calculation of the current data sequence and each of the case data sequences of each of the reference cases a similarity degree; from the case data sequences of each of the reference cases, select the case data sequence with the highest similarity to the current data sequence as a similar case data sequence in each of the reference cases, wherein The similar case data sequence corresponds to a similar case treatment policy; a multiple regression analysis is performed using the similar case data sequence and the similar case treatment policy to calculate a fitness function , wherein the strain number of the fitness function is a wound change; a parameter optimization algorithm is performed using the current data sequence and the fitness function to calculate an optimal treatment policy that maximizes the wound change and the An expected wound change value that should be optimally managed. The wound multiple sensing method of claim 1, further comprising: a first operation: using the current data sequence and the expected wound change value to estimate an expected data sequence, wherein the expected data sequence includes at least one size parameter; Second operation: use the expected data sequence and the fitness function to perform the parameter optimization algorithm to calculate the next best treatment policy that maximizes the wound change and the next best treatment policy corresponding to the next best treatment policy the expected wound change value, and then use the expected data sequence and the next expected wound change value to estimate the next expected data sequence; wherein, the next expected data sequence is used as the new expected data sequence to repeat the second execution The operation is performed until the at least one size parameter in the next expected data sequence is calculated to be 0. The wound multiple sensing method of claim 2, wherein the at least one size parameter is an expected wound length, an expected wound width and an expected wound depth. The wound multiple sensing method of claim 2, further comprising: accumulating the number of times the second operation is performed until the at least one size parameter in the next expected data sequence is calculated to be 0; The number of executions is incremented by 1 and multiplied by the period sensed to obtain an expected wound healing days. The wound multiple sensing method of claim 1, wherein the wound variation includes the following three dimensions: a wound length variation, a wound width variation, and a wound depth variation; and wherein maximizing the wound variation comprises: maximizing a wound volume and wherein the expected wound variation value includes an expected wound length variation value, an expected wound width variation value and an expected wound depth variation value. The wound multiple sensing method of claim 1, further comprising: using a color sensor in a wound machine to obtain the multiple wound observation data a wound tissue data; use a depth sensor in the wound machine to obtain a wound area surface area data in the plurality of wound observation data; use a temperature sensor in the wound machine to obtain the plurality of wound observation data A wound temperature data in . According to the wound multi-sensing method of claim 1, each of the reference cases further includes a characterization record, a qualitative questionnaire record and a care treatment record; wherein the case data sequence is established based on the characterization record; and where such similar case disposition policies are established based on the qualitative questionnaire records and the care disposition records. A wound multiple sensing system, comprising a processor for performing: reading a plurality of reference cases, wherein each of the reference cases includes a plurality of case data sequences corresponding to different time points; receiving a current wound of a wound Characterization, wherein the current wound representation includes a plurality of wound observation data; based on the plurality of wound observation data of the wound, a current data sequence is established; calculation of the current data sequence and each of the case data sequences of each of the reference cases a similarity degree; from the case data sequences of each of the reference cases, select the case data sequence with the highest similarity to the current data sequence as a similar case data sequence in each of the reference cases, wherein The similar case data sequence corresponds to a similar case handling policy; performing a multiple regression analysis using the similar case data series and the similar case treatment policies to calculate a fitness function, wherein the strain of the fitness function is a wound change; using the current data series and the adaptation The degree function performs a parameter optimization algorithm to calculate an optimal treatment policy that maximizes the wound change and an expected wound change value corresponding to the optimal treatment policy. According to the wound multi-sensing system of claim 8, the processor further performs: a first operation: using the current data sequence and the expected wound change value to estimate an expected data sequence, wherein the expected data sequence includes at least one size parameter; Second operation: use the expected data sequence and the fitness function to execute the parameter optimization algorithm to calculate the next best treatment policy that maximizes the wound change and the next best treatment policy corresponding to the next best treatment policy. wound change value, and then use the expected data sequence and the next expected wound change value to estimate the next expected data sequence; wherein, the next expected data sequence is used by the processor as the new expected data sequence to repeatedly execute the first The second operation is performed until the at least one size parameter in the next expected data sequence is calculated to be 0. The wound multi-sensing system of claim 9, wherein the at least one dimension parameter is an expected wound length, an expected wound width and an expected wound depth. According to the wound multiple sensing system of claim 9, the processor further executes: accumulating the number of times the second operation is performed until it is calculated that the at least one size parameter in the next expected data sequence is all 0; The number of times the second operation is performed is added by 1 and multiplied by the sensed period to obtain an expected number of days for wound healing. The wound multiple sensing system of claim 8, wherein the wound variation includes the following three dimensions: a wound length variation, a wound width variation, and a wound depth variation; and wherein maximizing the wound variation comprises: maximizing a wound volume and wherein the expected wound variation value includes an expected wound length variation value, an expected wound width variation value and an expected wound depth variation value. The wound multi-sensing system of claim 8 further includes a wound machine, the wound machine comprising: a color sensor for obtaining a wound tissue data among the plurality of wound observation data; a depth sensor for to obtain a wound area surface area data among the multiple wound observation data; a temperature sensor for obtaining a wound temperature data among the multiple wound observation data. The wound multi-sensing system of claim 8, wherein each of the reference cases further includes a characterization record, a qualitative questionnaire record, and a care treatment record; wherein the case data sequences are established based on the characterization record; and Wherein, the similar case disposal policies are established based on the qualitative questionnaire records and the care disposal records.

Images