Health care companion robot based on large model

Abstract

The present invention belongs to the field of robot technology. A health care companion robot based on a large model is provided, including a processing device, a camera device, a storage device, and a video playback device. The storage device stores a large model and a basic computer program, and the processing device calls and runs the basic computer program to make decisions using the large model. The health care companion robot of the present invention can automatically search for the video content that the elderly person likes and automatically play it, which can significantly improve the convenience of use for the elderly; it can also analyze the eye health of the elderly with eye diseases, and then determine whether it is appropriate to play the video content to the elderly. Only when the conclusion is appropriate will the screened video content be played to the elderly with eye diseases. Such a setting can avoid damage to the eyes of the elderly due to inappropriate video playback.

Description

Technical Field

[0001] The present invention relates to the field of robotics technology, and in particular to a health care companion robot based on a large model. Background Art

[0002] Health care companion robots are intelligent robots designed specifically for the elderly. They can provide a variety of services including life companionship, health monitoring, emergency calls, etc. Some health care companion robots are also equipped with audio-visual equipment, which can play videos and other consultations for the elderly, further enriching their lives.

[0003] However, when existing health care companion robots play videos for the elderly through audio-visual equipment, the elderly still need to operate manually, such as manually selecting or voice controlling the robot to play the content they want to watch. This is not conducive to improving the convenience of use for the elderly, and for the elderly with hand defects and language expression defects, this method obviously has obvious applicability defects.

[0004] Therefore, how to provide video content to the elderly more conveniently to enrich their lives is a technical problem that needs to be solved at present. Summary of the invention

[0005] In response to the above technical problems, the present invention provides a health care companion robot, electronic device, computer storage medium and computer program product based on a large model.

[0006] The present invention discloses a health care companion robot based on a large model, the health care companion robot comprises a processing device, a camera device, a storage device, and a video playing device, the storage device stores a large model and a basic computer program, the processing device calls and runs the basic computer program to make decisions using the large model; wherein the camera device is used to shoot an archive image of a target elderly person and a first live video of the target elderly person, and transmit the archive image and the first live video to the processing device; the processing device is used to process the received archive image and the first live video, and implement the following steps: perform character recognition on the archive image, extract archive information of the target elderly person, the archive information includes personal hobby information, eye medical information, and care remarks added by management personnel; according to the individual The method comprises the following steps: searching for a plurality of candidate video contents based on the eye hobby information, and sorting the candidate video contents from large to small according to the probability of preference; calling the large model to perform in-depth processing on the eye medical information and the nursing notes information, predicting a preliminary eye health recovery degree, and matching the preliminary eye health recovery degree to obtain the corresponding video playback duration; extracting the eye-related information of the target elderly from the first live video, using the large model to perform in-depth processing on the eye-related information, and predicting the eye health coefficient of the target elderly; calculating the target eye health recovery degree based on the eye health coefficient and the preliminary eye health recovery degree, and if the target eye health recovery degree is higher than a preset value, controlling the video playback device to play the sorted candidate video contents to the target elderly in sequence until the video playback duration is reached.

[0007] As an optional embodiment, the method of shooting the archival image of the target elderly person includes: when the health care companion robot in an idle state receives a new health care companion task, controlling the camera device to start up to shoot the archival image of the target elderly person displayed by the staff; the processing device is also used to: extract the facial image and the area where the target elderly person is located from the archival image, control the camera device to shoot a second live video in the area after walking to the area, identify the target elderly person from the second live video, and walk to the vicinity of the target elderly person to shoot the first live video of him.

[0008] As an optional embodiment, calling the big model to perform deep processing on the eye medical information and the nursing notes information to predict a preliminary eye health recovery degree includes: calling the big model in the storage device, and receiving the preliminary eye health recovery degree predicted by the big model as feedback; wherein, in response to the calling, the big model performs the following steps: retrieving historical medical information of the same type of eye diseases according to the eye medical information, and obtaining a historical dressing change information set according to statistics of the historical medical information, wherein the historical dressing change information set includes multiple historical dressing change information, and each of the historical dressing change information includes multiple historical dressing change time nodes and corresponding multiple historical eye health recovery degrees; performing semantic analysis on the nursing notes information, extracting the actual dressing change time nodes and the recorded actual eye health recovery degrees to form an actual dressing change information set; analyzing and calculating the overall deviation between the actual dressing change information set and the historical dressing change information set, and calculating the preliminary eye health recovery degree according to the overall deviation and the number of dressing changes in the actual dressing change information set; and feeding back the preliminary eye health recovery degree to the processing device.

[0009] As an optional embodiment, the eye-related information of the target elderly person is extracted from the first live video, the eye-related information is deeply processed using the large model, and the eye health coefficient of the target elderly person is predicted, including: extracting the eye movement characteristics and the hand intervention movement characteristics of the target elderly person from the first live video, the eye movement characteristics and the hand intervention movement characteristics of the eye constituting the eye-related information; using the large model to deeply process the eye movement characteristics and the hand intervention movement characteristics of the eye to predict a preliminary eye health coefficient of the target elderly person; extracting the sunlight exposure characteristics of the eyes of the target elderly person from the first live video, evaluating the sunlight exposure intensity based on the sunlight exposure characteristics, and determining the interference coefficient based on the sunlight exposure intensity; calculating the eye health coefficient of the target elderly person based on the preliminary eye health coefficient and the interference coefficient.

[0010] As an optional embodiment, the video playback device is controlled to play the sorted alternative video contents to the target elderly person in sequence, including: determining a viewing distance according to the target eye health recovery degree, controlling walking to the viewing distance, and controlling the video playback device to play the sorted alternative video contents to the target elderly person in sequence.

[0011] As an optional embodiment, the method of controlling the video playback device to play the sorted alternative video contents to the target elderly person in sequence also includes: extracting the head posture of the target elderly person from the first live video, and determining the playback posture of the video playback device according to the head posture; and controlling the video playback device to execute the playback posture so that it is the same as the head posture.

[0012] As an optional embodiment, before controlling the video playback device to play the sorted alternative video contents to the target elderly person in sequence, the method further includes: if the sunlight intensity is higher than an intensity threshold, controlling the video playback device to play a prompt message, wherein the prompt message is used to prompt the target elderly person to change his seating area.

[0013] The present invention also discloses an electronic device, which is applied to a large-model-based health care companion robot as described in any of the above items, and includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor.

[0014] The present invention also discloses a computer storage medium, which is applied to a large-model-based health care companion robot as described in any of the above items, and the computer-readable storage medium stores a computer program.

[0015] The present invention also discloses a computer program product, which is applied to a health care companion robot based on a large model as described in any of the above items, and the computer program product contains computer code.

[0016] The beneficial effects of the present invention are: 1) The health care companion robot does not require manual or voice operation by the elderly, but can automatically search for video content that the elderly like and play it automatically, which can significantly improve the convenience of use for the elderly with hand defects and language expression defects.

[0017] 2) We also conduct a special eye health analysis on elderly people with eye diseases to determine whether it is appropriate to play video content to the elderly. Only when the conclusion is appropriate will the screened video content be played to the elderly with eye diseases. This setting can avoid damage to the elderly's eyes caused by inappropriate video playback. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 It is a structural schematic diagram of a health care companion robot based on a large model disclosed in an embodiment of the present invention.

[0020] Figure 2 It is a schematic diagram of the processing flow of the processing device disclosed in the embodiment of the present invention.

[0021] Figure 3 It is a structural schematic diagram of an electronic device disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The following is a description of the implementation of the present application by specific specific embodiments. People familiar with the technology can easily understand other advantages and effects of the present application from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0023] In addition, the technical features involved in the different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0024] When existing health care companion robots play videos for the elderly through audio-visual equipment, the elderly still need to operate manually, such as manually selecting or voice controlling the robot to play the content they want to watch, which is not conducive to improving the convenience of use for the elderly. In addition, for the elderly with hand defects and language expression defects, this method also obviously has obvious applicability defects. Therefore, how to provide video content for the elderly more conveniently to enrich their lives is a technical problem that needs to be solved at present.

[0025] In response to the above technical issues, such as Figure 1 As shown, an embodiment of the present invention discloses a health care companion robot based on a large model, the health care companion robot includes a processing device, a camera device, a storage device, and a video playback device, the storage device stores a large model and a basic computer program, the processing device calls and runs the basic computer program to make decisions using the large model; wherein the camera device is used to capture an archive image of a target elderly person and a first live video of the target elderly person, and transmit the archive image and the first live video to the processing device; Figure 2As shown, the processing device is used to process the received archival image and the first live video, and implement the following steps: perform character recognition on the archival image to extract the archival information of the target elderly person, the archival information including personal hobby information, eye medical information, and care notes information added by management personnel; search for a number of candidate video contents based on the personal hobby information, and sort the candidate video contents from large to small according to the probability of preference; call the large model to perform in-depth processing on the eye medical information and the care notes information, predict the initial eye health recovery degree, and The initial eye health recovery degree is matched to obtain the corresponding video playback duration; the eye-related information of the target elderly person is extracted from the first live video, and the eye-related information is deeply processed using the large model to predict the eye health coefficient of the target elderly person; the target eye health recovery degree is calculated according to the eye health coefficient and the initial eye health recovery degree (the two can be directly multiplied), and if the target eye health recovery degree is higher than a preset value, the video playback device is controlled to play the sorted alternative video contents to the target elderly person in turn until the video playback duration is reached.

[0026] The health care companion robot in the present invention does not require the elderly to perform manual or voice operations, but can automatically search for and play the video content that the elderly like, which can significantly improve the convenience of use for the elderly with hand defects and language expression defects. In addition, the eye health of the elderly with eye diseases is analyzed in particular to determine whether it is appropriate to play the video content to the elderly. Only when the conclusion is appropriate will the screened video content be played to the elderly with eye diseases. This setting can avoid damage to the eyes of the elderly due to inappropriate video playback.

[0027] It should be noted that the search for a number of candidate video contents based on personal preference information can be performed using existing methods, and the present invention does not specifically limit this.

[0028] In addition, in addition to the above-mentioned processing equipment, camera equipment, storage equipment, and video playback equipment, the health care companion robot in the present invention should also have, for example, a walking mechanism, a detection and navigation device, an angle adjustment mechanism for the video playback device, etc., which will not be described in detail.

[0029] As an optional embodiment, the method of shooting the archival image of the target elderly person includes: when the health care companion robot in an idle state receives a new health care companion task, controlling the camera device to start up to shoot the archival image of the target elderly person displayed by the staff; the processing device is also used to: extract the facial image and the area where the target elderly person is located from the archival image, control the camera device to shoot a second live video in the area after walking to the area, identify the target elderly person from the second live video, and walk to the vicinity of the target elderly person to shoot the first live video of him.

[0030] In this embodiment, the health care companion robot of the present invention can be used in nursing homes, medical institutions or families, and a suitable number of health care companion robots can be deployed according to the actual needs of different usage scenarios. The health care companion robot can receive health care companion tasks issued by the staff in a variety of ways, such as sending health care companion tasks to the health care companion robot in an idle state through wireless methods such as Bluetooth and WiFi. The health care companion robot walks to the work area of ​​the staff, and the staff then shows the archival materials of the target elderly who need health care companionship to the health care companion robot. The health care companion robot controls the camera device to start shooting the archival materials of the target elderly displayed by the staff to obtain archival images, and the archival materials at least include the face image of the target elderly, the area (building, room number, specific geographical location, etc.), and the personal hobby information, eye medical information, and nursing notes added by the management personnel of the target elderly. Among them, personal hobby information refers to the interests and hobbies of the target elderly, based on which the above-mentioned alternative video content can be analyzed; eye medical information refers to the record of the target elderly receiving eye medical treatment, such as glaucoma, eye pterygium (especially pterygium, which is a common eye disease among the elderly), etc., and also includes the specific time of receiving eye medical treatment, the time of each dressing change, etc.; care notes information refers to the care points for the target elderly added by the management personnel, such as "control the length of video watching to 1 hour", "recent eye recovery is good / poor", etc.

[0031] At the same time, after extracting the facial image and the area where the target elderly person is located from the archival image, the processing device goes to the area on its own and shoots a second live video in the area to find and identify the target elderly person. After identifying the target elderly person, the processing device walks to the vicinity of the target elderly person to shoot the first live video of him or her.

[0032] As an optional embodiment, calling the big model to perform deep processing on the eye medical information and the nursing notes information to predict a preliminary eye health recovery degree includes: calling the big model in the storage device, and receiving the preliminary eye health recovery degree predicted by the big model as feedback; wherein, in response to the calling, the big model performs the following steps: retrieving historical medical information of the same type of eye diseases according to the eye medical information, and obtaining a historical dressing change information set according to statistics of the historical medical information, wherein the historical dressing change information set includes multiple historical dressing change information, and each of the historical dressing change information includes multiple historical dressing change time nodes and corresponding multiple historical eye health recovery degrees; performing semantic analysis on the nursing notes information, extracting the actual dressing change time nodes and the recorded actual eye health recovery degrees to form an actual dressing change information set; analyzing and calculating the overall deviation between the actual dressing change information set and the historical dressing change information set, and calculating the preliminary eye health recovery degree according to the overall deviation and the number of dressing changes in the actual dressing change information set; and feeding back the preliminary eye health recovery degree to the processing device.

[0033] In this embodiment, the processing device sends a scheduling request to the large model in the storage device, and the large model executes in response to the scheduling request: First, historical medical information of the same type of eye diseases is retrieved based on the eye medical information. For example, if the target elderly person undergoes pterygium surgery, historical medical information related to pterygium surgery is obtained from a database or other network resources. These historical medical information include the time when different patients undergo pterygium surgery and the subsequent dressing change time. Statistical analysis of the historical medical information can obtain a historical dressing change information set, which includes multiple historical dressing change information, each of which includes multiple historical dressing change time nodes and corresponding multiple historical eye health recovery degrees. The historical eye health recovery degree is recorded by medical staff when changing the patient's dressing or automatically evaluated based on the image of the surgical site. It should be noted that in order to ensure privacy, the patient's personal information needs to be deleted from the retrieved historical medical information.

[0034] Then, the above-mentioned nursing notes information is semantically analyzed to extract the actual dressing change time node and the actual recorded eye health recovery degree to form an actual dressing change information set. The actual dressing change information set represents the postoperative processing data actually performed by the target elderly. Next, the interval between the historical dressing change time node and the operation time of the corresponding number of dressing changes (i.e., the Nth dressing change) in multiple historical dressing change information, and the interval between adjacent historical dressing change time nodes can be averaged to obtain a standard dressing change time node set; then analyze and calculate the overall deviation between the actual dressing change information set and the standard dressing change time node set, where the standard dressing change time node set and the actual dressing change information set can be fitted into curves respectively, and then the overall deviation of the two curves is calculated.

[0035] Finally, the initial eye health recovery degree of the target elderly can be calculated comprehensively based on the overall deviation and the number of dressing changes in the actual dressing change information set. Among them, the initial eye health recovery degree = f1 (number of dressing changes) * f2 (overall deviation), f1, f2 can be a conversion formula or a comparison table. In other words, the corresponding eye health recovery value is first determined according to the number of dressing changes (which is positively correlated with the number of dressing changes), and then the corresponding adjustment coefficient is determined according to the overall deviation (which is negatively correlated with the overall deviation and is a value less than 1). The adjustment coefficient is multiplied by the eye health recovery value to obtain the initial eye health recovery degree. The large model feeds back the predicted initial eye health recovery degree to the processing device.

[0036] Among them, there is an obvious positive correlation between the video playback duration and the initial eye health recovery degree, that is, the higher the initial eye health recovery degree, the longer the video playback time can be provided to the target elderly people, and vice versa, the shorter the video playback time can be provided to the target elderly people to reduce the impact on their eye recovery.

[0037] It should be noted that the above-mentioned large model in the present invention refers to a general large model, which refers to a large-scale, powerful deep learning model that can handle a variety of different tasks and data types. It usually learns various features and patterns by training a large amount of data, thereby achieving fast and accurate processing of new data. Typical general large models are, for example, the GPT series based on Transformer, BERT, etc.

[0038] As an optional embodiment, the eye-related information of the target elderly person is extracted from the first live video, the eye-related information is deeply processed using the large model, and the eye health coefficient of the target elderly person is predicted, including: extracting the eye movement characteristics and the hand intervention movement characteristics of the target elderly person from the first live video, the eye movement characteristics and the hand intervention movement characteristics of the eye constituting the eye-related information; using the large model to deeply process the eye movement characteristics and the hand intervention movement characteristics of the eye to predict a preliminary eye health coefficient of the target elderly person; extracting the sunlight exposure characteristics of the eyes of the target elderly person from the first live video, evaluating the sunlight exposure intensity based on the sunlight exposure characteristics, and determining the interference coefficient based on the sunlight exposure intensity; calculating the eye health coefficient of the target elderly person based on the preliminary eye health coefficient and the interference coefficient.

[0039] In this embodiment, a preliminary analysis of the eye health recovery degree of the target elderly person can be achieved through the scheme of the aforementioned embodiment. The analysis method is actually based on the similarity of historical data, but there are large individual differences in eye recovery ability among different patients. The credibility of the preliminary eye health recovery degree predicted by the above scheme may deviate greatly from the actual situation of the target elderly person.

[0040] In view of the above situation, the present invention further controls the health care companion robot to extract the eye movement characteristics and hand intervention movement characteristics of the target elderly from the first live video, wherein the eye movement characteristics mainly refer to the number of times the target elderly blink and squint their eyes in the first unit time. The more frequent the blinking and squeezing, the more uncomfortable their eyes are, and the greater the probability of the corresponding poor postoperative recovery (worse than the aforementioned predicted initial eye health recovery); the hand intervention movement characteristics of the eyes mainly refer to the number of times the target elderly rub their eyes and squeeze eye drops and other eye medications in the second unit time (longer than the first unit time). The more frequent the rubbing and squeezing of eye drops, the more uncomfortable their eyes are, and the greater the probability of the corresponding poor postoperative recovery. Using a large model to deeply process the eye movement characteristics and the hand intervention movement characteristics of the eyes, the initial eye health coefficient of the target elderly can be predicted.

[0041] At the same time, the target elderly person's blinking, squeezing, rubbing eyes, etc. may also be affected by sunlight. For example, if the target elderly person sits next to a window, strong sunlight will irradiate the target elderly person's eyes, which will accelerate the occurrence of dry eyes, eye pain, itchy eyes, etc., and further aggravate the target elderly person's blinking, squeezing, rubbing eyes, etc. In this regard, the present invention further extracts the sunlight irradiation intensity of the target elderly person's eyes from the first live video (it can be determined according to the brightness of the sunlight irradiation area, and can also be determined according to the clarity of the light irradiation edge. The higher the clarity, the higher the sunlight irradiation intensity). The higher the sunlight irradiation intensity, the greater the corresponding interference coefficient. At this time, the preliminary eye health coefficient predicted by the aforementioned large model will be lower than the actual situation. It is necessary to use a larger interference coefficient to multiply the preliminary eye health coefficient of the target elderly person to appropriately increase it to be closer to the actual situation; otherwise, a smaller interference coefficient is used to adjust the preliminary eye health coefficient.

[0042] As an optional embodiment, the video playback device is controlled to play the sorted alternative video contents to the target elderly person in sequence, including: determining a viewing distance according to the target eye health recovery degree, controlling walking to the viewing distance, and controlling the video playback device to play the sorted alternative video contents to the target elderly person in sequence.

[0043] In this embodiment, since the target eye health recovery degree of the target elderly is different, it is necessary to determine the appropriate viewing distance for them. For example, if the target elderly has an eye problem that requires surgery, it is necessary to determine a gradually decreasing viewing distance according to the target eye health recovery degree. Specifically, the higher the target eye health recovery degree, the smaller the viewing distance, until the closest viewing distance is reached; the lower the target eye health recovery degree, the larger the viewing distance, until the farthest viewing distance is reached.

[0044] As an optional embodiment, the method of controlling the video playback device to play the sorted alternative video contents to the target elderly person in sequence also includes: extracting the head posture of the target elderly person from the first live video, and determining the playback posture of the video playback device according to the head posture; and controlling the video playback device to execute the playback posture so that it is the same as the head posture.

[0045] In this embodiment, the target elderly person may have a variety of head postures. To ensure that the target elderly person has a better viewing angle when watching the video content, the video playback device can be controlled to switch its playback posture to adapt to the head posture of the target elderly person. For example, if the target elderly person tilts his head to the left to watch, the video playback device is controlled to be offset to the right by a corresponding angle.

[0046] As an optional embodiment, before controlling the video playback device to play the sorted alternative video contents to the target elderly person in sequence, the method further includes: if the sunlight intensity is higher than an intensity threshold, controlling the video playback device to play a prompt message, wherein the prompt message is used to prompt the target elderly person to change his seating area.

[0047] In this embodiment, when it is found that the sunlight intensity in the area where the target elderly person is seated is too strong, playing video content for the target elderly person at this time is not conducive to the recovery of his eyes. Therefore, the video playback device is further controlled to play prompt information to the target elderly person to prompt him to change his seat area.

[0048] like Figure 3 As shown, an embodiment of the present invention further discloses an electronic device, which is applied to a large-model-based health care companion robot as described in any of the above items, comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor.

[0049] An embodiment of the present invention further discloses a computer storage medium, which is applied to a large-model-based health care companion robot as described in any of the above items, and the computer storage medium stores a computer program.

[0050] An embodiment of the present invention further discloses a computer program product, which is applied to a large-model-based health care companion robot as described in any of the above items, and the computer program product contains computer code.

[0051] The computer-readable storage medium described above may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the above. Alternatively, the computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media may include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above.

[0052] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0053] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0054] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A health care companion robot based on a large model, characterized by: The system comprises a processing device, a camera device, a storage device, and a video playback device. The storage device stores a large model and a basic computer program. The processing device uses the large model to make a decision by calling and running the basic computer program. The camera device is used to take an archive image of a target elderly person and a first live video of the target elderly person, and transmit the images to the processing device. The processing device is used to implement the following steps based on the archive image and the first live video: Perform character recognition on the archive images to extract the archive information of the target elderly, including personal hobby information, eye medical information, and care notes added by managers; A number of candidate video contents are obtained based on the personal preference information, and the candidate video contents are sorted from large to small according to the probability of preference; The big model is called to deeply process the eye medical information and nursing notes information to predict the initial eye health recovery degree, including: obtaining the historical dressing change information set corresponding to the same type of eye disease according to the eye medical information, obtaining the actual dressing change information set according to the nursing notes information, calculating the overall deviation of the two information sets, and obtaining the initial eye health recovery degree according to the overall deviation and the number of dressing changes in the actual dressing change information set; according to the initial eye health recovery degree, the corresponding video playback time is matched; Extracting eye-related information of the target elderly from the first live video, using a large model to deeply process the eye-related information, and predicting the eye health coefficient of the target elderly; the eye movement characteristics of the target elderly and the hand intervention movement characteristics on the eyes constitute the eye-related information; The target eye health recovery degree is calculated based on the eye health coefficient and the preliminary eye health recovery degree. If the target eye health recovery degree is higher than the preset value, the video playback device is controlled to play the sorted alternative video contents to the target elderly in sequence until the video playback time is reached.

2. A large-model health care companion robot according to claim 1, characterized in that: The method of shooting the archival image of the target elderly person includes: when the health care companion robot in an idle state receives a new health care companion task, controlling the camera device to start up to shoot the archival image of the target elderly person displayed by the staff; the processing device is also used to: extract the face image and the area where the target elderly person is located from the archival image, control the camera device to shoot a second live video in the area after walking to the area, identify the target elderly person from the second live video, and walk to the vicinity of the target elderly person to shoot the first live video of him.

3. A large-model-based health care companion robot according to claim 2, characterized in that: The calling of the big model to perform in-depth processing on the eye medical information and the nursing notes information, and predicting a preliminary eye health recovery degree, includes: calling the big model in the storage device, and receiving the preliminary eye health recovery degree predicted by the big model as feedback; wherein, in response to the calling, the big model performs the following steps: retrieving historical medical information of the same type of eye diseases according to the eye medical information, and obtaining a historical dressing change information set according to statistics of the historical medical information, wherein the historical dressing change information set includes multiple historical dressing change information, and each of the historical dressing change information includes multiple historical dressing change time nodes and corresponding multiple historical eye health recovery degrees; performing semantic analysis on the nursing notes information, extracting the actual dressing change time nodes and the recorded actual eye health recovery degrees, and forming an actual dressing change information set; and feeding back the preliminary eye health recovery degree to the processing device.

4. A large-model-based health care companion robot according to claim 3, characterized in that: The method of extracting eye-related information of the target elderly person from the first live video, deeply processing the eye-related information using a large model, and predicting the eye health coefficient of the target elderly person includes: extracting eye movement characteristics and hand intervention movement characteristics of the target elderly person from the first live video; deeply processing the eye movement characteristics and the hand intervention movement characteristics of the eyes using the large model to predict the preliminary eye health coefficient of the target elderly person; extracting sunlight exposure characteristics of the eyes of the target elderly person from the first live video, evaluating the sunlight exposure intensity based on the sunlight exposure characteristics, and determining the interference coefficient based on the sunlight exposure intensity; and calculating the eye health coefficient of the target elderly person based on the preliminary eye health coefficient and the interference coefficient.

5. A large-model-based health care companion robot according to claim 4, characterized in that: The method of controlling the video playback device to play the sorted candidate video contents to the target elderly in sequence includes: determining a viewing distance according to the target eye health recovery degree, controlling walking to the viewing distance, and controlling the video playback device to play the sorted candidate video contents to the target elderly in sequence.

6. A large-model-based health care companion robot according to claim 5, characterized in that: The method of controlling the video playback device to play the sorted alternative video contents to the target elderly person in sequence also includes: extracting the head posture of the target elderly person from the first live video, and determining the playback posture of the video playback device according to the head posture; and controlling the video playback device to execute the playback posture so that it is the same as the head posture.

7. A large-model-based health care companion robot according to claim 6, characterized in that: Before the video playback device is controlled to play the sorted candidate video contents to the target elderly in sequence, it also includes: if the sunlight intensity is higher than the intensity threshold, the video playback device is controlled to play prompt information, and the prompt information is used to prompt the target elderly to change the seating area.

8. An electronic device, characterized in that: A large-model-based health care companion robot as described in any one of claims 1 to 7 comprises: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor.

9. A computer storage medium, characterized in that: A health care companion robot based on a large model as described in any one of claims 1 to 7, wherein the computer storage medium stores a computer program.

10. A computer program product, characterized in that: A health care companion robot based on a large model as described in any one of claims 1 to 7, wherein the computer program product contains computer code.

Images