Devices, systems, and methods for tracking and evaluating wounds

Abstract

A method for tracking and evaluating a wound includes: identifying a wound from an image output by a camera; identifying a wound feature set based on the image; detecting a temperature of the wound from at least one thermal sensor; triggering at least one response if the wound temperature is outside the treatment temperature range; processing the wound feature set and the wound temperature using machine learning; determining an optimal wound healing trajectory based on the wound feature set and the wound temperature using machine learning; predicting at least one future wound condition based on the optimal wound healing trajectory using machine learning; and displaying the optimal wound healing trajectory and the at least one future wound condition on a display.

Description

Technical Field

[0001] The present invention relates generally to wound monitoring and, more particularly, to devices, systems and methods for objectively tracking and assessing wounds in a subject. Background Art

[0002] A subject may suffer from different types of wounds that may or may not require hospitalization. Changes in wound characteristics are a time-consuming and laborious task that many clinicians may have difficulty finding the time to adequately record. Various solutions have been proposed to automate the task of recording changes in wound characteristics. However, these solutions may not adequately consider wound temperature and its impact on the optimal wound healing trajectory. In addition, these solutions fail to provide a mechanism for shortening the optimal wound healing trajectory. Therefore, in the task of automating the task of recording changes in wound characteristics, wound temperature and a mechanism for shortening the optimal wound healing trajectory may need to be considered. Summary of the invention

[0003] In one aspect, a method for tracking and assessing a wound may include: identifying a wound from an image output by a camera; identifying a wound feature set based on the image; detecting a temperature of the wound from at least one thermal sensor; triggering at least one response if the wound temperature is outside a therapeutic temperature range; and processing the wound feature set and the wound temperature using machine learning. The method may also include: determining an optimal wound healing trajectory based on the wound feature set and the wound temperature using machine learning; predicting at least one future wound condition based on the optimal wound healing trajectory using machine learning; and displaying the optimal wound healing trajectory and at least one future wound condition on a display.

[0004] In another aspect, a system for tracking and assessing wounds may include: a display; a camera configured to output image data; at least one thermal sensor configured to output temperature data; a memory containing a machine-readable medium including machine-executable code having instructions stored thereon; and a controller communicatively coupled to the display, the camera, the at least one thermal sensor, and the memory. The controller includes one or more processors and is configured to execute the machine-executable code to cause the controller to: identify a wound and a wound feature set from the image data; detect a temperature of the wound from the at least one thermal sensor; trigger at least one response if the wound temperature is outside a therapeutic temperature range; process the wound feature set and the wound temperature using machine learning; determine an optimal wound healing trajectory based on the wound feature set and the wound temperature using machine learning; predict at least one future wound condition based on the optimal wound healing trajectory using machine learning; and display the optimal wound healing trajectory and the at least one future wound condition on the display.

[0005] On the other hand, a method for tracking and evaluating burn wounds may include: identifying a burn wound from an image output by a camera; capturing a set of burn wound features based on the image; classifying the burn wound based on sensor data from one or more sensors; processing the burn wound features and the sensor data with a machine learning algorithm to determine an optimal burn wound healing trajectory, wherein the machine learning algorithm is trained with a burn wound statistical library; predicting at least one future burn wound condition based on the optimal wound healing trajectory with the machine learning algorithm; and displaying the optimal burn wound healing trajectory and at least one future burn wound condition on a display.

[0006] In another aspect, a non-transitory, processor-readable storage medium may include one or more programming instructions stored thereon that, when executed, cause a processing device to perform a method for tracking and assessing wounds and / or burn wounds.

[0007] These and other features and characteristics of the present technology, as well as methods of operation and function of the related structural elements and combinations of parts and economies of manufacture will become more apparent when considering the following description and the appended claims with reference to the accompanying drawings, all of which form a part of this specification, wherein like reference numerals indicate corresponding parts in the various figures. However, it is to be expressly understood that the drawings are for illustration and description purposes only and are not intended as a definition of limitations on the scope of the invention. As used in the specification and claims, the singular forms "a", "an" and "the" include plural referents unless the context clearly dictates otherwise. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] The embodiments set forth in the drawings are illustrative and exemplary in nature and are not intended to limit the subject matter defined by the claims. Figure 1 The following detailed description of illustrative embodiments may be understood when read together with the accompanying drawings, in which like structures are represented by like reference numerals, and in which:

[0009] Figure 1 schematically depicts an example system for tracking and assessing wounds according to one or more embodiments shown and described herein;

[0010] Figure 2 schematically depicts an example smart device and heater for use in the implementation of a method for tracking and assessing a wound according to one or more embodiments shown and described herein;

[0011] Figure 3 schematically depicts an example heater for use in the implementation of a method for tracking and assessing a wound according to one or more embodiments shown and described herein;

[0012] Figure 4A flowchart depicting an example method of tracking and assessing a wound according to one or more embodiments shown and described herein; and

[0013] Figure 5 Depicted is a flow diagram of an example method of tracking and assessing a burn wound in accordance with one or more embodiments shown and described herein. DETAILED DESCRIPTION

[0014] The present invention relates to devices, systems and methods for objectively tracking and evaluating wounds. In particular, a wound is identified from an image, and a wound feature set is identified based on the image. The temperature of the wound is detected in real time, and a response is triggered if the wound temperature is outside a therapeutic temperature range. In particular, the response may include activating a heater to warm the wound. The wound feature set and the wound temperature are input into a machine learning algorithm for processing using machine learning, and an optimal wound healing trajectory is determined. Real-time temperature detection improves existing wound monitoring systems because decisions affecting wound healing can be made more quickly. The ability to warm the wound using a heater improves existing wound monitoring systems by providing a mechanism to shorten the optimal wound healing trajectory.

[0015] As described above, a wound is identified from an image, and a wound feature set is identified based on the image. The temperature of the wound is detected in real time, and a response is triggered if the wound temperature is outside the treatment temperature range. This response may include activating a heater to warm the wound. The ability to use a heater to warm the wound improves existing wound monitoring systems by providing a mechanism to shorten the optimal wound healing trajectory. The wound feature set and the wound temperature are input into a machine learning algorithm for processing using machine learning, and the optimal wound healing trajectory is determined. The machine learning algorithm is trained on a wound statistics library. The use of a wound statistics library improves existing wound monitoring systems because the accuracy of determining the optimal wound healing trajectory increases over time. In addition, since tracking and evaluation are based on objective factors, the devices, systems, and methods disclosed herein objectively track and evaluate wounds. For example, the tracking and evaluation described herein may be based on wound location, size, shape, color, depth, temperature, humidity, etc.

[0016] Turning now to the accompanying drawings, Figure 1A wound tracking and assessment system 100 is shown according to one or more embodiments of the present invention. The system 100 includes a smart device 106 having a camera 108 and a thermal sensor 110, which can be used (e.g., arranged, positioned, etc.) to image a wound 104 on an object 102. The smart device 106 can be communicatively coupled to a wired or wireless network 112, and images and / or temperature data from the camera 108 and the thermal sensor 110 can be sent through the network 112. As used herein, the term "temperature data" can also include humidity data. The network 112 can be communicatively coupled to other smart devices, such as a heater 114 and / or one or more temperature probes 116. The network 112 can also be communicatively coupled to other computing devices, such as a computer 118 and / or one or more displays 120. The network 112 can also be communicatively coupled to a remote server 122 and a database 124. Thus, the smart device 106 including the camera 108 and the thermal sensor 110 , the heater 114 , one or more temperature probes 116 , the computer 118 and one or more displays 120 , and the server 122 and database 124 may each be communicatively coupled to one another via the network 112 .

[0017] The wound 104 on the object 102 can be any kind of wound, including a wound on the skin or a wound that can be seen with the camera 108 or the thermal sensor 110. For example, the wound 104 can be a penetrating wound, such as a puncture wound, a surgical wound / incision, a thermal, chemical or electrical burn, a bite and stab wound, and / or a gunshot wound. The wound 104 can also be from blunt force trauma, such as an abrasion, a laceration, or a skin tear. The wound 104 can also be a closed wound, such as a contusion, a blister, a seroma, a hematoma, or a crush injury. In addition, the wound 104 can also be a skin ulcer. The wound 104 can be of different sizes or colors and can be at any stage of the healing process. In some cases, the wound 104 may not be visible to the naked eye (e.g., localized), but can be imaged using the devices described herein (e.g., a subcutaneous wound that can be captured / detected via infrared imaging, etc.). In addition, although the wound 104 is Figures 1 to 3 1 as being located on the arm of subject 102, but it should be understood that wound 104 may be located anywhere on the subject's body without departing from the scope of the present invention.

[0018] The smart device 106 may be a mobile phone, tablet computer, or other suitable computing device. In some embodiments, the smart device 106 is a controller for controlling the operation of the system 100. In this regard, the smart device 106 includes a processor and a non-transitory storage module that are communicatively coupled to each other, wherein the non-transitory storage module of the smart device 106 includes computer-readable and executable instructions that can be executed by the processor of the smart device 106. In addition, the smart device 106 may include a touch screen display 107 for outputting data for viewing by the subject 102 and / or the medical team. The smart device 106 may send image data from the camera 108, temperature and / or humidity data from the thermal sensor 110, and other data related to the wound via the network 112 (e.g., via communication hardware (not described)). The smart device 106 may also include a graphical user interface (GUI) for inputting data for the subject 102 and / or the medical team and for controlling the operation of various components of the system 100. That is, initialization of the system 100 may be achieved via the smart device 106.

[0019] The camera 108 is configured to capture images and / or videos of the wound 104, and may include multiple cameras that work together or independently to capture images from various angles. The thermal sensor 110 may be included as part of the camera 108 or separated from the camera 108 and placed outside it. Therefore, the camera 108 and / or the thermal sensor 110 are configured to capture images within any electromagnetic frequency range, including the visible spectrum and the infrared (IR) spectrum. The camera 108 may include various types of cameras and / or optical sensors. For example, the camera 108 may include various types of optical sensors, including RGB, LiDAR optical sensors, or a combination thereof. The thermal sensor 110 may include a thermal imager or an infrared sensor, such as (e.g., a camera module provided by Teledyne FLIR (Wilsonville, Oregon)). In addition, the thermal sensor 110 may be configured to determine surface moisture. In an embodiment, the camera 108 may be configured to capture still images and / or videos. The camera 108 may capture still images, multiple images within a predetermined time period, and / or videos. Thermal sensor 110 is configured to detect temperature and / or surface humidity of a wound in real time. As used herein, "real time" refers to the ability of thermal sensor 110 (or other device) to provide instantaneous and continuous updates of temperature and / or surface humidity as the temperature and / or surface humidity changes. As used herein, "real time" may include a response within a given time delay, which may vary from sub-seconds to seconds.

[0020] The network 112 may be configured as any wired and / or wireless network, such as, for example, a wide area network, the Internet, an intranet, etc. The network may also be configured with hardware, including any device capable of sending and / or receiving data via the wired or wireless network 112. For example, the hardware of the network 112 may include a chipset (e.g., an antenna, a processor, machine-readable instructions, etc.) for communicating via a wired and / or wireless computer network, such as Wireless Fidelity (Wi-Fi), WiMax, Bluetooth, etc.

[0021] The heater 114 is configured to supply at least one type of heat to the wound 104. For example, the heater 114 can supply radiant heat and / or convective heat to the wound 104. Devices that supply radiant heat, such as the heater 114 in some embodiments, are generally configured to emit heat in the form of electromagnetic waves that travel directly and heat objects and surfaces without requiring a medium such as air to carry the heat. Devices that supply convective heat, such as the heater 114 in some embodiments, are generally configured to transfer heat through the movement of a fluid caused by differences in density and temperature within the fluid. For example, the heater 114 may include a heating element and a fan that quickly distributes warm air. In some embodiments, such as Figure 2 In the embodiment shown in , the heater 114 can be mounted to a bracket that can be placed close to the wound 104. However, the heater 114 described herein is not limited to being mounted to a bracket. That is, in other embodiments, the heater 114 can be mounted to another component, or can be hand-held, so that the heater 114 is allowed to be positioned close to the wound 104. For example, the heater 114 can be mounted to a bed or chair on which the subject 102 can lie or sit. The heater 114 can also be attached to an articulated arm or boom mounted on the ceiling or wall, which allows the heater 114 to be positioned in different positions and directions relative to the wound 104 of the subject 102. Alternatively, the heater 114 can also be mounted directly to the ceiling or wall.

[0022] refer to Figure 3In some embodiments, heater 214 may be mounted to the body of subject 102. That is, heater 214 may be configured as an electrically heated sleeve, wrap, blanket, sheath, etc. that may be mounted on the body of subject 102 at the location of wound 104. Similar to heater 114, body-mounted heater 214 may be configured to provide at least one of radiant heat and / or convective heat to wound 104. In some embodiments, body-mounted heater 214 may also be configured with a moisturizing device (not shown) that may be activated to provide moisture to wound 104, for example, by applying water or a disinfected / sterile liquid suitable for application to wound 104. Heaters 114, 214 may be powered by alternating current (AC) or direct current (DC).

[0023] In some embodiments, the heater 114, 214 may be provided with a chipset (e.g., antenna, processor, machine-readable instructions, etc.) to communicate via the network 112, thereby allowing remote control of the heater 114, 214. For example, the smart device 106 and / or the computer 118 may be used to control the heater 114, 214.

[0024] One or more temperature probes 116 may be used in conjunction with thermal sensor 110 to obtain an accurate temperature measurement of wound 104. In addition, one or more temperature probes 116 may be configured to detect surface moisture of wound 104. In this regard, Figure 2 As shown, one or more temperature probes 116 are typically placed at the outer edge of the wound 104, as close to the wound bed as possible, but not directly on the wound bed, to monitor and measure the temperature and / or humidity of the wound bed in real time. The outer edge of the wound 104 is generally defined as a path that surrounds, surrounds, or outlines the two-dimensional shape of the wound on the skin of the subject 102. The determination of the outer edge of the wound can be based on a visual inspection of the color change between the typical wound color, tone, or texture and the natural tone of the skin of the subject 102. The determination of the outer edge of the wound can also be based on a visual inspection of the height change between the wound 104 and the surrounding skin of the subject 102, where the wound 104 can be convex or concave relative to the normal plane extending from the skin of the subject 102. As discussed in further detail below, the system 100 itself can be configured to determine the outer edge of the wound 104 and notify the placement of one or more temperature probes 116 at the outer edge of the wound 104. In addition, the one or more temperature probes 116 can also be configured to have a chipset (e.g., an antenna, a processor, machine-readable instructions, etc.) to transmit temperature and / or humidity readings via the network 112.

[0025] The computer 118 may be any suitable computing device that generally includes a processor and a non-transitory storage module that are communicatively coupled to each other, wherein the non-transitory storage module of the computer 118 includes computer-readable and executable instructions that can be executed by the processor of the computer 118. Thus, in some embodiments, the computer 118 is a controller that controls the operation of the system 100. In some embodiments, the computer 118 may additionally or alternatively control the system 100 from the smart device 106. The computer 118 also includes one or more displays 120 for outputting data for viewing by the subject and / or the medical team. The computer 118 may also include a graphical user interface (GUI) for the subject 102 and / or the medical team to input data and for controlling the operation of various components of the system 100. That is, initialization of the system 100 is also achieved via the smart device 106.

[0026] In an embodiment, the system 100 may be configured to establish a network connection to communicate with one or more remote servers 122. For example, the smart device 106 or the computer 118 may communicate wirelessly or via a wired connection with one or more remote servers 122. In this regard, data from the camera 108 and / or the thermal sensor 110 may be remotely streamed on a device / display associated with the remote server 122.

[0027] In an embodiment, the system 100 may be configured to establish a network connection to communicate with one or more databases 124. For example, the smart device 106 or computer 118 may communicate with the database 124 wirelessly or via a wired connection. In this regard, data from the camera 108 and / or thermal sensor 110 may be stored remotely on the database 124. Additionally, depending on the nature of the algorithm stored in the memory of the smart device 106 and / or computer 118, the algorithm may generate an output, and such output may be stored remotely on the database 124.

[0028] Now refer to Figure 4 A method for tracking and assessing a wound 104 using the system 100 discussed above is described. In some embodiments, the various steps of the method may be performed by a smart device 106, a computer 118, and / or a remote server 122. The various steps may be performed on separate computing devices and control systems, or all steps may be performed by a single processor. For example, processing of image and temperature data may be performed on the smart device 106, the computer 118, or the remote server 122. As another example, certain portions of the processing that require a heavier CPU load may occur on the remote server 122. As an additional example, a scaled-down version of the machine learning algorithm may be run locally on the smart device 106 or the computer 118.

[0029] At block 400, one or more temperature probes 116 are applied to the skin of the subject 102, near the outer edge of the wound 104, to help detect the temperature and / or humidity of the wound 104. The one or more temperature probes 116 send temperature data in real time via the network 112, and the temperature and / or humidity are received by the smart device 106, the computer 118, and / or the server 122 of the system 100. In addition, the temperature data from the one or more temperature probes 116 can be stored in the memory of the smart device 106 and / or the memory of the computer 118. The use of one or more temperature probes 116 advantageously provides more accurate temperature and / or humidity monitoring than using a thermal sensor (e.g., thermal sensor 110) alone. In addition, real-time temperature and / or humidity detection by one or more temperature probes 116 (or by thermal sensor 110) described herein improves existing wound monitoring systems because decisions affecting wound healing can be made more quickly. For example, as described in further detail below, a decision to activate the heater 114, 214 can be made more quickly to ensure that the wound 104 remains at a temperature within the therapeutic temperature range.

[0030] Figure 4 The method continues at block 402, where the thermal sensor 110 is directed at the wound 104 to detect a constant wound temperature and / or humidity. The thermal sensor 110 transmits temperature data, including humidity data, in real time via the network 112, which can be received by the smart device 106, the computer 118, and / or the server 122 of the system 100. In addition, the temperature data from the thermal sensor 110 can be stored in a memory of the smart device 106 and / or a memory of the computer 118.

[0031] At block 404, wound temperature data from the one or more temperature probes 116 and thermal sensors 110 is provided to a service provider / medical team of subject 102. For example, the wound temperature data may be sent to smart device 106 or computer 118 so that the medical team of subject 102 may view the wound temperature data on display 107 of smart device 106 or display 120 of computer 118. The wound temperature data may also be stored in a memory of smart device 106 and / or a memory of computer 118 for subsequent retrieval, manipulation, and / or management.

[0032] Box 404 also includes providing the wound feature set to the service provider / medical team of the subject 102. To obtain the wound feature set, the system 100 identifies the wound 104 from the image output by the camera 108. Thus, the subject 102 or a member of the subject's medical team can capture an image of the wound 104 using the camera 108 of the smart device 106, and the image can be received by the processor of the smart device 106. The image can be sent via the network 112 in addition or alternatively and received by the processor of the computer 118. The image can include a color image and / or can include temperature data output by the thermal sensor 110 and / or one or more temperature probes 116. The system 100 can process the image to identify the wound boundary. That is, the processor of the smart device 106 or the computer 118 can process the image to identify the wound boundary. This can include determining the location of pixels representing the outer edge of the wound 104. In some embodiments, a color change threshold from a typical wound color, tone, or texture to the natural tone of the subject 102 skin can represent the boundary of the wound 104. In other embodiments, changes in height or depth detected by a lidar or similar sensor may be used to detect the boundaries of the wound 104. In other embodiments, various computer vision processing techniques may be used to identify the wound 104 and its boundaries, as known to those skilled in the art. In addition, temperature and humidity measurements from the thermal sensor 110 and / or one or more temperature probes 116 may be mapped to the wound boundary to determine which measurements output from the thermal sensor 110 and / or one or more temperature probes 116 are within the wound boundary.

[0033] Once the wound 104 is identified by the image output by the camera 108, the system 100 obtains a set of wound features within the wound boundary based on the image. The wound feature set may include features such as temperature and / or humidity (such as output by the thermal sensor 110 and / or one or more temperature probes 116), size, shape, color and / or depth. The system 100 may determine the temperature and / or humidity features based on the output of the thermal sensor 110 and / or one or more temperature probes 116, and may determine the temperature and / or humidity features in combination with the wound boundary information from the color image. The system 100 may also determine the size and shape (e.g., circular, elliptical, irregular or linear) of the wound 104 based on the image data using one or more algorithms known to those skilled in the art. The system 100 may also determine the color of the wound 104, including various other optical features of the image of the wound 104 output by the camera 108. Different colors may indicate or be associated with different stages of healing of the wound 104. In addition, different colors may indicate inflammation or infection of the wound 104. For example, red may indicate granulation tissue, yellow may indicate molting, and black may indicate eschar or necrotic tissue. The system 100 can also determine the depth of the wound 104, including superficial, partial thickness, or full-thickness involvement, based on the image data output by the camera 108, and the camera 108 can include a lidar sensor or other suitable sensor. The system 100 can also determine the location of the wound 104 on the body of the object 102. For example, the location can be determined as a bony protrusion or a high friction area. The system 100 can also determine the edge characteristics of the wound 104, such as maceration, erosion, or curling. The system can also determine the presence, type, and amount of wound drainage or exudate, which can be serous (transparent), hemorrhagic (bloody), serum (transparent with blood), purulent (yellow, green, or brown), or mixed.

[0034] The wound feature set obtained as described above is provided to a service provider / medical team at block 404. For example, the wound feature set is sent to the smart device 106 or computer 118 so that the medical team of the subject 102 can view the set of wound features on the display 107 of the smart device 106 or the display 120 of the computer 118. The wound feature set may also be stored in the memory of the smart device 106 and / or the memory of the computer 118 for subsequent retrieval, manipulation, and / or management. For example, the wound feature set may be used in subsequent steps to determine an optimal wound healing trajectory, as discussed in further detail below.

[0035] As described above, the system 100 itself can be configured to determine the outer edge or border of the wound 104 and inform the placement of one or more temperature probes 116 at the outer edge / border of the wound 104. In this regard, as described above, the system 100 identifies the outer edge / border of the wound 104 and the wound feature set from the image output by the camera 108. After the outer edge / border of the wound 104 and the wound feature set have been obtained, the system 100 can mark the image with the recommended location for placing the one or more temperature probes 116 in consideration of the outer edge / border and the wound feature set, and display the marked image on the smart device 106 or computer 118. For example, the outer edge of the wound 104 can be identified, and the placement location of the one or more temperature probes 116 on the outer edge can be selected based on the proximity of the placement location to the identified features of the wound. For example, the placement location can be selected on the outer edge of the wound based on the maximum diameter, width, or depth of the identified wound, as long as the placement location does not overlap with the identified infected area to prevent other disturbances / infections in the area. As another example, the system 100 can also recommend a specific number of temperature probes 116 to be placed on the outer edge of the wound based on the size of the identified wound and the wound feature set. In this regard, the system 100 is configured to strategically recommend the placement of one or more temperature probes 116 in a manner that does not hinder or is unlikely to hinder the healing of the wound 104.

[0036] At box 406, if the wound temperature is outside the treatment temperature range, the system 100 triggers at least one response. In some embodiments, at least one response includes sending an alarm if the wound temperature is outside the treatment temperature range. The alarm can be sent to at least one of the smart device 106, the computer 118, or the server 122. If the wound is hypothermic and the wound temperature is outside the treatment temperature range, at least one response can also include warming the wound 104 to the treatment temperature range at box 408. When the system 100 commands the heater 114 (or heater 214) to turn on and remain on until the wound 104 is detected to be warmed to the normal temperature within the treatment area by the thermal sensor 110 and / or one or more temperature probes 116, the warming of the wound 104 is achieved. The ability to warm the wound 104 within the treatment temperature range using the heater 114 or 214 improves existing wound monitoring systems by providing a mechanism to shorten the optimal wound healing trajectory. Therefore, keeping the wound in the appropriate temperature zone is critical to the continued healing of the wound 104. Since faster healing leads to shorter hospital stays, there is less chance of complications during the healing process.

[0037] Block 406 may also include triggering at least one response if the wound moisture is outside the therapeutic moisture range (e.g., the wound is too dry). In some embodiments, at least one response includes sending an alarm if the wound moisture is outside the therapeutic moisture range. The alarm may be a text-based message indicating that the wound needs to be moisturized and that water or a disinfected / sterile liquid should be applied to the wound. In embodiments using a body-mounted heater 214, at least one response may also include activating an atomizer of the heater 214 to supply moisture to the wound 104.

[0038] At block 410, once the wound temperature reaches the therapeutic range, the system 100 instructs the heater 114 (or heater 214) to turn off. After the heater 114 (or heater 214) is turned off, the thermal sensor 110 and / or the one or more temperature probes 116 continue to detect the wound temperature. At least one response may also include automatically recording the wound feature set and the wound temperature / humidity in an electronic medical record (EMR) at block 412. The EMR may be included as part of the database 124, or may be a separate database configured to send and / or receive data via the network 112.

[0039] The smart device 106, computer 118, and / or server 122 may maintain a machine learning algorithm or model that has been trained using the wound statistics library to determine an optimal wound healing trajectory, predict at least one future wound condition, and / or output wound healing recommendations. After training the machine learning algorithm, in operation, the camera 108, thermal sensor 110, and / or one or more temperature probes 116 capture data (e.g., images, temperature, and / or humidity) of the wound 104 of the subject 102. The collected data is input into the trained machine learning algorithm at block 414, which outputs a determined optimal wound healing trajectory, at least one future wound condition predicted based on the optimal wound healing trajectory, and / or a wound healing recommendation at block 416. Additional details regarding the determined optimal wound healing trajectory, the predicted at least one future wound condition, and / or the wound healing recommendation will be described in more detail below.

[0040] During the training of the machine learning algorithm, a subject volunteer or a service provider / medical team may provide various wound images and wound data (e.g., temperature and / or humidity). The wound images and wound data may be pre-existing or may be captured using the smart device 106 (e.g., using the camera 108 and the thermal sensor 110). The captured data may be labeled to generate training data based on the specific healing stage of the wound at the time of capturing the image or other sensor data (e.g., the wound has been formed for one week, the wound has been formed for one month). The labeled captured data may include a wound statistics library, which includes various wound images at various stages of healing and a set of related wound features for each wound image at each stage of healing. Therefore, the wound statistics library may include, but is not limited to, at least one of similar wound features, similar wound temperature / humidity, and the stage of wound healing. The wound statistics library may also include, but is not limited to, at least one of subject demographics, subject comorbidities, or therapies and procedures that affect wound healing. The wound statistics library may be stored in a database 124. Therefore, a variety of training data may be captured to train the machine learning algorithm without overfitting the model to a narrow set of training data. The smart device 106, computer 118, and / or server 122 can then train a machine learning algorithm with the wound statistics library to determine an optimal wound healing trajectory, predict at least one future wound condition, and / or output a wound healing recommendation, as disclosed in further detail below. Using the wound statistics library to train a machine learning algorithm improves existing wound monitoring systems because the accuracy of the determined optimal wound healing trajectory and the predicted at least one future wound condition improves over time. That is, as the system 100 tracks and evaluates additional wounds, data obtained from these additional wounds is added to the wound statistics library, thereby increasing the amount of data that the machine learning algorithm can make determinations and / or predictions about.

[0041] When the camera 108, the thermal sensor 110, and / or the one or more temperature probes 116 capture wound images and other sensor data, the captured wound images and sensor data can be sent to the smart device 106, the computer 118, and / or the server 122. The smart device 106, the computer 118, and / or the server 122 can receive images from the camera 108 and data from the thermal sensor 110 and / or the one or more temperature probes 116, and can train and maintain a machine learning algorithm to determine an optimal wound healing trajectory, at least one future wound condition, and / or wound healing recommendations based on the received images and sensor data, as disclosed herein. During training, the smart device 106, the computer 118, and / or the server 122 can receive training data (e.g., a wound statistics library), and can train the machine learning algorithm based on the received training data using, for example, supervised learning techniques. After the machine learning algorithm is trained with the wound statistics library, during real-time operation, the smart device 106, computer 118, and / or server 122 receives real-time data from the camera 108, thermal sensor 110, and / or one or more temperature probes 116 (e.g., wound feature sets), and inputs the received data (e.g., wound feature sets) into the trained algorithm for processing using the trained algorithm. The smart device 106, computer 118, and / or server 122 can determine an optimal wound healing trajectory, predict at least one future wound condition, and / or output wound healing recommendations based on the received data.

[0042] As used herein, "machine learning algorithm" refers to a set of mathematical and statistical techniques and procedures that enable a computer, such as smart device 106, computer 118, and / or server 122, to automatically learn and improve its performance on a specific task or problem without being explicitly programmed. The term "machine learning algorithm" as used herein may refer to various types, including but not limited to supervised learning, unsupervised learning, and / or reinforcement learning. Specific examples of machine learning algorithms used herein may include but are not limited to linear regression, decision trees, support vector machines, k-means clustering, and / or neural networks.

[0043] As used herein, "real-time operation" refers to the operation of the system 100 to determine an optimal wound healing trajectory, predict at least one future wound condition, and / or output wound healing recommendations after the machine learning model has been trained. As used herein, "real-time data" refers to data received from the camera 108, the thermal sensor 110, and / or the one or more temperature probes 116 during real-time operation.

[0044] The above-mentioned machine learning algorithm can be included as a program module in the memory of the smart device 106, the computer 118 and / or the server 122. The program module can be in the form of an operating system, an application module, and other program modules stored in the memory. Such a program may include, but is not limited to, routines, subroutines, programs, objects, components, data structures, etc. that perform specific tasks or execute specific data types. For example, the program module may include programming for causing the processor of the smart device 106 and / or the computer 118 to receive sensor data from the camera 108, the thermal sensor 110, and / or one or more temperature probes 116. In addition, the program module may include programming for causing the processor of the smart device 106 and / or the computer 118 to pre-process the sensor data from the camera 108, the thermal sensor 110, and / or one or more temperature probes 116 before inputting the data into the machine learning algorithm for processing using machine learning. The pre-processing of the sensor data may include combining different types of data into a single training example to be input into the machine learning algorithm. The pre-processing of the sensor data may also include formatting the data, filtering the data, cropping the image, and the like. In addition, the program modules may include programming for causing the processor of the smart device 106 and / or computer 118 to train a machine learning algorithm using, for example, supervised learning techniques. In addition, the program modules may include programming for causing the processor of the smart device 106 and / or computer 118 to determine an optimal wound healing trajectory, predict at least one future wound condition, and / or output wound healing recommendations.

[0045] The wound feature set and the wound temperature data are input into the machine learning algorithm for processing using machine learning, and the machine learning algorithm is configured to determine an optimal wound healing trajectory at block 414. In other words, the machine learning algorithm is configured to output various information related to the optimal wound healing trajectory, including, but not limited to, the healing stage of the wound 104, the remaining time for the wound 104 to heal, and the therapeutic temperature range for optimal healing of the wound 104. The optimal wound healing trajectory is also sent to the smart device 106 or computer 118 so that the medical team of the subject 102 can view the optimal wound healing trajectory on the display 107 of the smart device 106 and / or the display 120 of the computer 118. In this regard, the machine learning algorithm may include as an output a representation of the optimal wound healing trajectory in the form of a text-based message.

[0046] Based on the optimal wound healing trajectory, the machine learning algorithm also predicts at least one future wound condition. The at least one future wound condition can also be described as a set of future wound characteristics. Therefore, the at least one future wound condition includes, but is not limited to, the future temperature, future humidity, future size, future shape, future color, and / or future depth of the wound 104. The at least one future wound condition is also sent to the smart device 106 or the computer 118 so that the medical team of the subject 102 can view the at least one future wound condition on the display 107 of the smart device 106 and / or the display 120 of the computer 118. In this regard, the machine learning algorithm can include as an output a representation of the at least one future wound condition in the form of a 2D or 3D computer-generated image.

[0047] At box 416, the machine learning algorithm is also configured to output a wound healing recommendation as a result of the analysis of the wound feature set and the wound temperature data. The wound healing recommendation includes, but is not limited to, best practices for wound care or risks associated with wound healing. The wound healing recommendation is also sent to the smart device 106 or computer 118 so that the medical team of the subject 102 can view the wound healing recommendation on the display 107 of the smart device 106 and / or the display 120 of the computer 118. In this regard, the machine learning algorithm can include as an output a representation of the wound healing recommendation in the form of a text-based message.

[0048] As described above, best practices for wound care and risks to wound healing are suggested at box 416, which the service provider / medical team can implement to assist in complying with the optimal wound healing trajectory determined at box 414. In this regard, the wound feature set and wound temperature change over time, and the system 100 continues to track and assess the wound 104 as these changes occur. That is, the system 100 is configured to adjust to the current state of the wound. For example, the wound 104 may ultimately enter a healing range corresponding to a different treatment temperature range than when the initial tracking and assessment of the wound 104 began. The healing range can be determined based on the temperature detected by the thermal sensor 110 and one or more temperature probes 116 reaching a temperature change threshold, which can be different for different types of wounds.

[0049] Therefore, at block 418, if the wound temperature is outside the new or healing treatment temperature range, the system 100 proceeds to trigger at least one response. In some embodiments, the at least one response includes sending an alert if the wound temperature is outside the healing treatment temperature range. The alert may be sent to at least one of the smart device 106, the computer 118, or the server 122. The at least one response may also include warming the wound to the healing treatment temperature range using the heater 114 (or heater 214) as described above with respect to block 406.

[0050] At block 420, at least one response may additionally include automatically updating the wound feature set and wound temperature in the EMR of subject 102. In other words, the analysis of wound 104 may be updated in close time to assist in meeting optimal healing time. As used herein, "close time" means as close to real time as possible. In the context of updating the EMR, the EMR may incorporate the updated analysis as quickly as possible.

[0051] Over time, the system 100 can determine that the wound 104 has progressed to the end of the optimal healing trajectory based on the wound feature set and the wound temperature. That is, the system 100 can determine that the wound feature set and the wound temperature have reached a threshold indicating that the optimal healing trajectory has been met. Therefore, at block 422, the system 100 collects the clinical course and outcomes of the wound 104 and reports them to national, local, and institutional wound registries.

[0052] Now refer to Figure 5 A method for tracking and assessing a burn wound using the system 100 discussed above is described. Figure 5 The method shown is similar to Figure 4 The method shown is similar except that Figure 5 The method shown is specific to burn wounds. In this regard, for the sake of brevity, reference will be made to Figure 4 Method to describe Figure 5 In some embodiments, various steps of the method may be performed by the smart device 106, the computer 118, and / or the remote server 122. Various steps may be performed on separate computing devices and control systems, or all steps may be performed by a single processor.

[0053] At block 500, a subject (e.g., subject 102) suffers a burn injury in the field. For example, the burn injury may be a thermal burn, a chemical burn, or an electrical burn.

[0054] At block 502, a burn wound feature set is captured using the smart device 106. To obtain the burn wound feature set, the system 100 identifies a burn wound (e.g., wound 104) from the image output by the camera 108. The identification of the burn wound is similar to the above description of Figure 4 The burn wound identification discussed in block 404 of the illustrated method is performed in a substantially similar manner. Thus, the processor of the smart device 106 or computer 118 can process the image to identify the burn wound boundary based on pixel location, color change threshold, height or depth change, and / or any other suitable computer vision processing technique. Once the burn wound is identified in the image output by the camera 108, the system 100 obtains a burn wound feature set within the burn wound boundary based on the image. The burn wound feature set is similar to the burn wound feature set described above with respect to Figure 4 The wound characteristic set described in block 404 of the illustrated method is obtained in a substantially similar manner. Thus, the burn wound characteristic set may include temperature, moisture, size, shape, color, and / or depth.

[0055] At block 504, the burn wound feature set thus obtained is provided to a service agency, such as a local emergency medical service (EMS). For example, the burn wound feature set is sent to the smart device 106 or the computer 118 so that the EMS can view the burn wound feature set on the display 107 of the smart device 106 or the display 120 of the computer 118. The burn wound feature set can also be stored in the memory of the smart device 106 and / or the memory of the computer 118 for subsequent retrieval, operation and / or management.

[0056] At block 506, the local EMS classifies the burn wound using the smart device 106. Classification of the burn wound is based on sensor data from one or more sensors (e.g., camera 108, which may be a depth camera, or thermal sensor 110, which may be a thermal imager). Classification is performed so that the initial extent of the burn injury can be determined prior to emergency room (ER) admission.

[0057] At block 508, the burn wound feature set and the sensor data are input into a machine learning algorithm for processing using machine learning, and an optimal burn wound healing trajectory is suggested via the machine learning algorithm. In other words, the burn wound feature set and the sensor data are processed by a machine learning algorithm as implemented by the smart device 106 and / or the computer 118 to determine an optimal burn wound healing trajectory. The optimal burn wound healing trajectory is sent to the smart device 106 or the computer 118 so that the medical team of the subject 102 can view the optimal burn wound healing trajectory on the display 107 of the smart device 106 and / or the display 120 of the computer 118. Figure 4 The method shown is similar, and the burn wound feature set and sensor data are also processed using a machine learning algorithm to predict at least one future burn wound condition based on the optimal wound healing trajectory. The at least one future burn wound condition is also displayed on the display 107 of the smart device 106 and / or the display 120 of the computer 118. Figure 4 As discussed in the method shown, the machine learning algorithm is trained using a burn wound statistical library. Similar to the wound statistical library discussed above, the burn wound statistical library may include, but is not limited to, similar burn wound characteristics, similar burn wound temperature / humidity, and the stage of burn wound healing. The burn wound statistical library may also include, but is not limited to, subject demographics, subject comorbidities, or therapies and procedures that affect burn wound healing.

[0058] At block 510, the burn wound feature set and the sensor data set are additionally processed, wherein the burn wound feature set and the sensor data set are input into a machine learning algorithm for processing using machine learning, and the machine learning algorithm outputs a burn wound healing recommendation. The burn wound healing recommendation may include, but is not limited to, at least one of best practices for burn wound care and / or risks associated with burn wound healing.

[0059] In some instances, the machine learning algorithm can output a burn wound healing recommendation, i.e., the best practice for burn wound care is to perform a skin graft surgery. In the case where a skin graft is recommended, at box 512, the burn wound is evaluated using system 100 after surgery. That is, the skin graft wound is evaluated using system 100. This postoperative evaluation may include repeating box 502 to capture a set of skin graft wound features. This postoperative evaluation also includes temperature and / or humidity evaluation via thermal sensor 110 (e.g., thermal imager) and / or one or more temperature probes 116. Therefore, the thermal sensor and / or one or more temperature probes 116 can be used as described in boxes 400 and 402 above to detect the temperature and / or humidity of the skin graft from one or more sensors in real time.

[0060] Once one or more sensors (e.g., thermal sensor 110 and / or one or more temperature probes 116) are detecting the real-time temperature, at block 514, if the skin graft temperature is outside the therapeutic temperature range, the system 100 can trigger at least one response. Figure 4 As discussed in the method of, if the skin graft temperature is outside the therapeutic temperature range, triggering at least one response may include sending an alarm. If the skin graft temperature is outside the therapeutic temperature range, triggering at least one response may also include warming the skin graft to the therapeutic temperature range. As described above, warming the skin graft is performed by heater 114 (or heater 214).

[0061] At block 516, triggering at least one response may also include identifying potential early skin graft failure based on the temperature assessment and communicating the potential early skin graft failure to the subject's service provider / medical team. That is, the machine learning algorithm may identify potential early skin graft failure based on the temperature detected by one or more sensors (e.g., thermal sensor 110 and / or one or more temperature probes 116).

[0062] At block 518, triggering at least one response may additionally include the machine learning algorithm suggesting an intervention for responding to potential early skin graft failure, such that the intervention may be captured by the subject's provider / medical team's records. For example, the intervention may be entered and stored in the subject's electronic medical record (EMR).

[0063] At block 520, triggering at least one response may also include updating an analysis of the skin graft wound within a proximate time to assist in meeting an optimal healing time.

[0064] Over time, the system 100 can determine that the skin graft has progressed to the end of the optimal healing trajectory. That is, the system 100 can determine that a set of skin graft wound characteristics and temperatures have reached a threshold indicating that the optimal healing trajectory has been met. Therefore, at block 522, the system 100 collects the clinical course and outcomes of the skin graft wound and reports them to national, local, and institutional burn wound registries.

[0065] It should be noted that the terms "substantially" and "approximately" may be used herein to represent the inherent degree of uncertainty that may be attributed to any quantitative comparison, value, measurement, or other representation. These terms are also utilized herein to represent the degree by which a quantitative representation may vary from a stated reference without resulting in a change in the basic function of the subject matter at issue.

[0066] Although specific embodiments have been illustrated and described herein, it should be understood that various other changes and modifications may be made without departing from the spirit and scope of the claimed subject matter. In addition, although various aspects of the claimed subject matter have been described herein, these aspects need not be used in combination. Therefore, it is intended that the appended claims cover all such changes and modifications within the scope of the claimed subject matter.

[0067] It should now be understood that the embodiments described herein are directed to methods for tracking and evaluating wounds. As described above, a wound is identified from an image, and a wound feature set is identified based on the image. The temperature and / or humidity of the wound is detected in real time, and a response is triggered if the wound temperature is outside the treatment temperature range and / or treatment humidity range. The real-time temperature detection described herein improves existing wound monitoring systems because decisions affecting wound healing can be made more quickly. Such a response may include activating a heater to warm the wound. The ability to warm the wound using a heater improves existing wound monitoring systems by providing a mechanism to shorten the optimal wound healing trajectory. The wound feature set and the wound temperature are input into a machine learning algorithm for processing using machine learning, and the optimal wound healing trajectory is determined. The machine learning algorithm is trained on a wound statistics library. The use of a wound statistics library improves existing wound monitoring systems because the accuracy of determining the optimal wound healing trajectory increases over time. In addition, the embodiments disclosed herein objectively track and evaluate wounds because the tracking and evaluation are based on objective factors. For example, the tracking and evaluation described herein can be based on wound location, size, shape, color, depth, temperature, humidity, etc.

[0068] Further aspects of the invention are provided by the subject matter of the following clauses.

[0069] A method for tracking and assessing a wound, the method comprising: identifying a wound from an image output by a camera; identifying a wound feature set based on the image; detecting the temperature of the wound from at least one thermal sensor; triggering at least one response if the wound temperature is outside a therapeutic temperature range; processing the wound feature set and the wound temperature using machine learning; determining an optimal wound healing trajectory based on the wound feature set and the wound temperature using machine learning; predicting at least one future wound condition based on the optimal wound healing trajectory using machine learning; and displaying the optimal wound healing trajectory and the at least one future wound condition on a display.

[0070] A method as in any preceding clause, wherein at least one of the at least one thermal sensor is a thermal imager.

[0071] A method according to any preceding clause, wherein at least one of the at least one thermal sensor is a temperature probe placed around an outer edge of the wound.

[0072] A method as in any preceding clause, wherein triggering the at least one response comprises sending an alarm if the wound temperature is outside the therapeutic temperature range.

[0073] A method as in any preceding clause, wherein triggering the at least one response comprises warming the wound to the therapeutic temperature range if the wound temperature is outside the therapeutic temperature range.

[0074] A method as in any preceding clause, wherein triggering the at least one response comprises recording the set of wound characteristics and the wound temperature in an electronic medical record.

[0075] A method according to any preceding clause, further comprising training the machine learning algorithm with a wound statistics library.

[0076] A method as in any preceding clause, wherein the wound statistics library comprises at least one of similar wound characteristics, similar wound temperatures, wound healing stages, subject demographics, subject comorbidities, or therapies and procedures that affect wound healing.

[0077] A method as in any preceding clause, further comprising outputting a wound healing recommendation as a result of analyzing the wound feature set and wound temperature using machine learning.

[0078] A method as in any preceding clause, wherein the wound healing recommendation comprises at least one of: best practices for care of the wound or risks associated with healing of the wound.

[0079] A method as in any preceding clause, wherein detecting the temperature of the wound from at least one thermal sensor is performed in real time.

[0080] A method as in any preceding clause, further comprising detecting moisture of the wound from the at least one thermal sensor.

[0081] A system for tracking and assessing wounds, the system comprising: a display; a camera configured to output image data; at least one thermal sensor configured to output temperature data; a memory containing a machine-readable medium, the machine-readable medium including machine-executable code having instructions stored thereon; a controller communicatively coupled to the display, the camera, the at least one thermal sensor and the memory, the controller comprising one or more processors and configured to execute the machine-executable code so that the controller: identifies a wound and a wound feature set from the image data; detects a temperature of the wound from the at least one thermal sensor; triggers at least one response if the wound temperature is outside a therapeutic temperature range; processes the wound feature set and the wound temperature using machine learning; determines an optimal wound healing trajectory based on the wound feature set and the wound temperature using machine learning; predicts at least one future wound condition based on the optimal wound healing trajectory using machine learning; and displays the optimal wound healing trajectory and the at least one future wound condition on the display.

[0082] A system according to any preceding clause, wherein the at least one thermal sensor is a thermal imager.

[0083] A system according to any preceding clause, wherein the at least one thermal sensor is one or more temperature probes placed around the periphery of the wound.

[0084] The system of any preceding clause, further comprising a heater configured to supply at least one type of heat to the wound, wherein the controller is communicatively coupled to the heater.

[0085] A system according to any preceding clause, wherein the at least one type of heat is radiant heat or convective heat.

[0086] A system according to any preceding clause, wherein the heater is an electric heating cuff placed over the wound.

[0087] A system according to any preceding clause, wherein the controller is further configured to activate the heater until the wound temperature reaches the therapeutic temperature range.

[0088] A system as described in any preceding clause, wherein detecting the temperature of the wound from the at least one thermal sensor is performed in real time.

[0089] A method for tracking and evaluating burn wounds, the method comprising: identifying a burn wound from an image output by a camera; capturing a burn wound feature set based on the image; classifying the burn wound based on sensor data from one or more sensors; processing the burn wound features and sensor data with a machine learning algorithm to determine an optimal burn wound healing trajectory, wherein the machine learning algorithm is trained with a burn wound statistical library; predicting at least one future burn wound condition based on the optimal wound healing trajectory with the machine learning algorithm; and displaying the optimal burn wound healing trajectory and the at least one future burn wound condition on a display.

[0090] A method as in any preceding clause, wherein the one or more sensors comprise at least one of a depth camera or a thermal imager.

[0091] The method of any preceding clause, wherein the burn wound statistical library comprises at least one of similar burn wound characteristics, stages of burn wound healing, subject demographics, subject comorbidities, or therapies and procedures that affect burn wound healing.

[0092] A method as in any preceding clause, wherein processing the burn wound characteristics and sensor data comprises inputting the burn wound characteristics and sensor data into the machine learning algorithm and outputting a burn wound healing recommendation.

[0093] A method as in any preceding clause, wherein the burn wound healing recommendations comprise at least one of: best care practices for care of the burn wound or risks associated with healing of the burn wound.

[0094] A method according to any preceding clause, wherein the burn wound is a skin graft and the method further comprises detecting in real time the temperature of the skin graft from the one or more sensors and triggering at least one response if the skin graft temperature is outside a therapeutic temperature range.

[0095] A method as in any preceding clause, wherein triggering the at least one response comprises sending an alarm if the skin graft temperature is outside the therapeutic temperature range, or warming the skin graft to the therapeutic temperature range if the skin graft temperature is outside the therapeutic temperature range.

[0096] A method as in any preceding clause, further comprising reporting at least one of the burn wound characteristics, sensor data, or optimal burn wound healing trajectory to one or more burn wound registries.

[0097] A non-transitory, processor-readable storage medium comprising: one or more programming instructions stored thereon, wherein the one or more programming instructions, when executed, cause a processing device to perform a method according to any one of the above clauses.

Claims

1. A method for tracking and assessing a wound, the method comprising: identifying wounds from images output by a camera; identifying a wound feature set based on the image; detecting a temperature of the wound from at least one thermal sensor; triggering at least one response if the wound temperature is outside of a therapeutic temperature range; Processing the wound feature set and the wound temperature using machine learning; determining an optimal wound healing trajectory based on the wound feature set and the wound temperature using machine learning; using machine learning to predict at least one future wound condition based on the optimal wound healing trajectory; as well as The optimal wound healing trajectory and the at least one future wound condition are displayed on a display. 2 . The method of claim 1 , wherein at least one of the at least one thermal sensor is a thermal imager.

3. The method of claim 1 or 2, wherein at least one of the at least one thermal sensor is a temperature probe placed around the outer edge of the wound.

4. The method of any of the preceding claims, wherein triggering the at least one response comprises sending an alarm if the wound temperature is outside the therapeutic temperature range.

5. The method of any of the preceding claims, wherein triggering the at least one response comprises warming the wound to the therapeutic temperature range if the wound temperature is outside the therapeutic temperature range.

6. The method of any of the preceding claims, wherein triggering the at least one response comprises recording the set of wound characteristics and the wound temperature in an electronic medical record.

7. The method according to any of the preceding claims, further comprising training a machine learning algorithm with a wound statistics library.

8. The method of claim 7, wherein the wound statistics library includes at least one of similar wound characteristics, similar wound temperatures, wound healing stages, subject demographics, subject comorbidities, or therapies and procedures that affect wound healing.

9. The method of any one of the preceding claims, further comprising outputting a wound healing recommendation as a result of analyzing the wound feature set and wound temperature using the machine learning.

10. The method of claim 9, wherein the wound healing recommendation comprises at least one of: best practices for care of the wound or risks associated with healing of the wound.

Images