Device and method for predicting early dementia in elderly person by using artificial intelligence

Abstract

According to one embodiment of the present invention, a device for predicting early dementia in an elderly person by using artificial intelligence comprises: an image data collection unit on which a subject draws an additional picture over an image on which stimulus symbols are printed, and which collects image data; an analysis unit for performing, on the image data, machine learning using a pre-stored artificial neural network model, and analyzing the dementia prediction degree; and an output unit for outputting dementia prediction information about the subject's degree of dementia analyzed by the analysis unit.

Description

Device and method for predicting early dementia in the elderly using artificial intelligence

[0001] The present invention relates to an apparatus and method for predicting early dementia in the elderly using artificial intelligence, and more specifically, to an apparatus and method for predicting whether the elderly have early dementia by using WZT (Wartegg-Zeichentest) images and deep learning analysis.

[0002] With increasing social interest in dementia, research aimed at identifying its causes and proving related risk factors is on the rise. Dementia can be a brain disease associated with aging. In this context, dementia can cause progressive memory impairment. Furthermore, dementia can lead to changes in an individual's personality or a decline in cognitive abilities.

[0003] Dementia can make daily life difficult, and even healthy individuals can develop it due to continuous brain function impairment. To prevent dementia in advance, it may be necessary to analyze the causes of dementia and related risk factors.

[0004] Generally, the progression of dementia can be confirmed through medical methods such as brain imaging or biomarkers. Additionally, for example, symptoms of dementia can be alleviated or worsened depending on physical activity, such as regular exercise or sleep.

[0005] Looking at the incidence of dementia, there is an increasing trend centered on the elderly as the population ages. Although the prevalence of dementia varies by country, it is estimated that more than 150 million people worldwide will develop dementia by 2050, and the massive socioeconomic cost of treatment is expected to be around $250 billion.

[0006] The need for a preventive approach to dementia is drawing attention, and predictions regarding the outcomes of prevention suggest that if the onset of dementia is delayed by about 5 years, the prevalence rate will be reduced by half, and if it is delayed by 3 years, the cost of dementia will be reduced by 1 / 3.

[0007] Dementia is diagnosed based on the judgment of a professional conducting clinical evaluations, such as a history of cognitive symptoms, physical examinations, paper cognitive assessments, and questionnaires. Early detection of dementia in the elderly is crucial, and the Wartegg-Zeichentest (WZT) makes it easier to provide early health information about dementia to seniors. The WZT method provides technology that can support the development of new, efficient, and accessible methods to help detect the early stages of dementia. The characteristics of WZT drawings reveal unconscious attitudes toward the artist's inner world and their unique relationships. WZT drawings encourage unconscious, honest, and natural expression while simultaneously incorporating specific patterns depending on the individual's disposition.

[0008] However, while such WZT dementia prediction is receiving significant attention in the natural language processing and social science communities, evaluating the surveys and projective methods for dementia assessment requires the participation of qualified professionals. There are drawbacks, such as the fact that it takes a considerable amount of time for a group of experts specializing in early-stage dementia to analyze and derive results, as well as significant costs and the effort involved in the complex evaluation of expert opinions.

[0009] Accordingly, the device and method for predicting early dementia in the elderly using artificial intelligence according to one embodiment of the present invention can significantly reduce the social costs of the expert group that diagnoses dementia by easily predicting dementia through the results of a deep learning automatic image inspection that can test for dementia.

[0010] An early dementia prediction device for the elderly utilizing artificial intelligence according to one embodiment of the present invention is characterized by comprising: an image data collection unit in which an examiner additionally draws on an image on which stimulus symbols are printed and collects image data; an analysis unit in which machine learning is performed using an artificial neural network model stored in the image data and the degree of dementia prediction is analyzed; and an output unit in which dementia prediction information regarding the degree of dementia of the examiner analyzed by the analysis unit is output.

[0011] An early dementia prediction device for the elderly utilizing artificial intelligence according to one embodiment of the present invention may be characterized in that the images collected by the data collection unit are composed of the WZT (Wartegg Zeichen Test).

[0012] The analysis unit of the device for predicting early dementia in the elderly using artificial intelligence according to one embodiment of the present invention may be characterized by being composed of a Canny Edge Detection method that divides the pixels of the image data into different intervals and verifies the boundaries of the image data.

[0013] The analysis unit of the device for predicting early dementia in the elderly using artificial intelligence according to one embodiment of the present invention may be characterized by using a Super-Resolution Generative Adversarial Network (SRGAN) that converts the image of the image data into a high-resolution image.

[0014] The analysis unit of the device for predicting early dementia in the elderly using artificial intelligence according to one embodiment of the present invention may be characterized by learning the image data using a Convolutional Neural Network (CNN) technique.

[0015] Here, the CNN may be characterized by being composed of CNN layers including a convolution layer, an activation layer, a pooling layer, and a fully connected layer.

[0016] In addition, the activation function used in the above CNN layer may be characterized by using the ReLU (Rectified Linear Unit) function and the Softmax function.

[0017] The analysis unit of the device for predicting early dementia in the elderly using artificial intelligence according to one embodiment of the present invention derives a learning result based on a softmax function, derives a probability of normal cognitive function and a probability of abnormal cognitive function based on the softmax function, and selects the probability of normal cognitive function and the probability of abnormal cognitive function as the final result, wherein if the probability of normal cognitive function is high, it is determined that the cognitive function is normal, and if the probability of abnormal cognitive function is high, it is determined that the cognitive function is abnormal.

[0018] The analysis unit of the device for predicting early dementia in the elderly using artificial intelligence according to one embodiment of the present invention may be characterized by deriving a learning result in which the probability of normal cognitive function and the probability of abnormal cognitive function are divided and a linear function is applied to a layer processing the probability of normal cognitive function and the probability of abnormal cognitive function, and if the resulting number is negative, it is classified as abnormal cognitive function, and if it is positive, it is classified as normal cognitive function.

[0019] A method for predicting early dementia in an elderly person using artificial intelligence according to one embodiment of the present invention may be characterized by comprising: an image data collection step in which an examiner additionally draws on an image on which stimulus symbols are printed and collects image data; a dementia prediction analysis step in which machine learning is performed using an artificial neural network model stored in the image data collected in the data collection step and the degree of dementia prediction is analyzed; and an output step in which dementia prediction information regarding the degree of dementia of the examiner analyzed from the dementia prediction analysis step is output.

[0020] A method for predicting early dementia in the elderly using artificial intelligence according to one embodiment of the present invention may be characterized in that the images collected in the data collection step are composed of the Wartegg Zeichen Test (WZT).

[0021] The dementia prediction analysis step of the method for predicting early dementia in the elderly using artificial intelligence according to one embodiment of the present invention may be characterized by being composed of a Canny Edge Detection method that divides the pixels of the image data into different intervals and verifies the boundaries of the image data.

[0022] The dementia prediction analysis step of the method for predicting early dementia in the elderly using artificial intelligence according to one embodiment of the present invention may be characterized by using a Super-Resolution Generative Adversarial Network (SRGAN) that converts the image of the image data into a high-resolution image.

[0023] In the dementia prediction analysis step of the method for predicting early dementia in the elderly using artificial intelligence according to one embodiment of the present invention, the image data may be trained using a Convolutional Neural Network (CNN) technique.

[0024] A method for predicting early dementia in an elderly person using artificial intelligence according to another embodiment of the present invention may be characterized by comprising: an image data collection unit that collects image data by applying a Convolutional Neural Network (CNN) technique to an image on which a stimulus symbol is printed and an examiner additionally draws a picture on the image; an analysis unit that performs machine learning on the image data using either a Canny Edge Detection method or a Super-Resolution Generative Adversarial Network (SRGAN) and analyzes the degree of dementia prediction; and an output unit that outputs dementia prediction information regarding the degree of dementia of the examiner analyzed by the analysis unit.

[0025] The purpose of the present invention is to provide a device and method that can significantly reduce the social costs of the expert group determining dementia by using artificial intelligence to predict early dementia in the elderly according to one embodiment of the present invention, as dementia can be easily predicted through the results of a deep learning automatic image inspection that can test for dementia.

[0026] The effects of the present invention are not limited to the technical problems mentioned above, and other unmentioned effects will be clearly understood by those skilled in the art from the description below.

[0027] A brief description of each drawing is provided to help to better understand the drawings cited in the detailed description of the invention.

[0028] FIG. 1 is a diagram illustrating an early dementia prediction device for the elderly using artificial intelligence according to an embodiment of the present invention.

[0029] FIG. 2 is a diagram illustrating image data for dementia testing related to early dementia prediction according to one embodiment of the present invention.

[0030] FIG. 3 is a diagram illustrating the process of deriving learning results based on a Canny Edge Detection model according to an embodiment of the present invention.

[0031] FIG. 4 is a diagram illustrating the process of deriving learning results based on an SRGAN model according to an embodiment of the present invention.

[0032] FIG. 5 is a diagram illustrating training with a CNN, which is a pre-stored artificial neural network model, in an early dementia prediction device for the elderly according to one embodiment of the present invention.

[0033] FIG. 6 is a diagram illustrating a procedure for classifying dementia prediction information using a CNN, which is an artificial neural network model stored in an early dementia prediction device for the elderly according to an embodiment of the present invention.

[0034] FIG. 7 is a diagram illustrating a method for predicting early dementia in the elderly using artificial intelligence according to an embodiment of the present invention.

[0035] Hereinafter, embodiments according to the present invention will be described with reference to the accompanying drawings. It should be noted that in assigning reference numerals to the components of each drawing, the same components are given the same reference numeral whenever possible, even if they are shown in different drawings. Furthermore, in describing the embodiments of the present invention, if it is determined that a detailed description of related known components or functions would hinder understanding of the embodiments of the present invention, such detailed description is omitted. Additionally, while embodiments of the present invention will be described below, the technical concept of the present invention is not limited or restricted thereto and can be modified and implemented in various ways by those skilled in the art.

[0036] Furthermore, the terms used herein are for the purpose of describing embodiments and are not intended to limit or / or restrict the disclosed invention. Singular expressions include plural expressions unless the context clearly indicates otherwise.

[0037] In this specification, terms such as “comprising,” “having,” or “having” are intended to indicate the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and do not preclude the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0038] Additionally, throughout the specification, when a part is described as being "connected" to another part, this includes not only cases where they are "directly connected" but also cases where they are "indirectly connected" with other elements in between. Terms including ordinal numbers, such as "first," "second," etc., used in this specification may be used to describe various components, but the components are not limited by these terms.

[0039] Below, embodiments of the present invention are described in detail with reference to the attached drawings so that those skilled in the art can easily implement the invention. Additionally, parts of the drawings that are irrelevant to the description are omitted to clearly explain the invention.

[0040] FIG. 1 is a diagram illustrating an early dementia prediction device for the elderly using artificial intelligence according to an embodiment of the present invention.

[0041] First, for the convenience of explanation, the device for predicting early dementia in the elderly using artificial intelligence according to the present invention will be referred to as a prediction device below.

[0042] Below, we will examine the internal configuration of the prediction device that performs important functions for the implementation of the present invention and the functions of each component.

[0043] As illustrated in FIG. 1, a prediction device according to one embodiment of the present invention includes a data collection unit (110), an analysis unit (120), an output unit (130), a communication unit (140), and a control unit (150). According to one embodiment of the present invention, the collection unit (110), the analysis unit (120), the output unit (130), the communication unit (140), and the control unit (150) may be program modules that communicate with an external system, at least in part. Such program modules may be included in the prediction device in the form of an operating system, an application program module, or other program modules, and may be physically stored in various known memory devices. Additionally, such program modules may be stored in a remote memory device capable of communicating with the early dementia prediction device. Meanwhile, such program modules encompass, but are not limited to, routines, subroutines, programs, objects, components, data structures, etc., that perform specific tasks or execute specific abstract data types as described below according to the present invention.

[0044] Meanwhile, although the prediction device has been described as above, this description is exemplary, and it is obvious to those skilled in the art that at least some of the components or functions of the early dementia prediction device may be realized or included within an external system (not shown) as needed.

[0045] First, a data collection unit (110) according to one embodiment of the present invention performs the role of collecting image data input by an examiner in correspondence with a picture on which a stimulus symbol is printed.

[0046] Specifically, the data collection unit (110) uses the WZT (Wartegg Zeichen Test), shows a picture with printed stimulus symbols to the examiner, has the examiner draw the stimulus symbols and the picture together, analyzes the features and meanings in the drawn picture, and collects the analyzed data as image data. Meanwhile, the WZT is a test in which the ability to recognize and respond to test stimuli corresponds to behavior in a social environment, and the WZT belongs to the category of "projection" tests and is classified based on performance.

[0047] As shown in Fig. 2, WZT may be the result of a projection test utilizing various types of images on eight stimulus images with a 4*4cm black border. The images are eight stimulus images, each with different meanings and interpretations, and can be used to identify the personality of the examiner.

[0048] The stimulus images of the WZT are interpreted as follows. The first stimulus image represents a central dot. This "central dot" expresses the meaning of self-experience. The second stimulus image represents "movement" floating in the air; it is unique and positioned in the upper left corner of the space to emphasize the characteristics of movement, thereby expressing imagination. The third stimulus image represents "ascent," starting low from the left and rising towards the right, expressing a strong inner goal orientation. The fourth stimulus image is a small square positioned in the upper right corner, representing a sense of "heaviness." In psychological counseling, it expresses mental burden and difficulty as it effectively induces inspiration. The fifth stimulus image features two contrasting lines, expressing "tension," "hostility," and "aggression." The extension and simultaneous contrast of the lines serve to clarify the inner self, implying the overcoming of tension and the suppression of formative turmoil. The sixth stimulus image, with its separated horizontal and numerical lines, is an expression that unconsciously demands "combination." The semicircle drawn with a dotted line in the bottom right of the seventh stimulus image represents delicacy and expresses sensitive sensibility. Finally, the arched curve of the eighth stimulus image signifies a 'sense of security'.

[0049] Meanwhile, in the present invention, the components that can be included as image data are not necessarily limited to using the above-mentioned WGT, and it can be assumed that any device capable of collecting various image data within the scope of achieving the purpose of the present invention may be employed.

[0050] Referring to FIG. 3, the analysis unit (S200) of the present invention divides the pixels of an image into different intervals and checks whether they are boundaries. Here, the analysis unit (200) of the present invention uses the Canny Edge Detection method, which is an edge detection algorithm that detects boundaries in an image. In addition, the Canny Edge Detection method detects edges through several steps to increase the continuity and accuracy of the boundaries. The figure on the left of FIG. 3 (S210) uses color and black-and-white data, but the figure on the right (S220) applies the Canny Edge Detection method, extracts the outline of the figure, recognizes the shape of an object as an outline, and marks the area.

[0051] Referring to FIG. 4, the analysis unit (300) of the present invention utilizes a learned artificial neural network model, and can be configured to learn by restoring a low-resolution picture to a high-resolution picture based on SRGAN (Super-resolution GAN).

[0052] Specifically, a prediction device according to one embodiment of the present invention can output image-enhanced image data by applying a style conversion function in an analysis unit (300) to image data based on an image drawn by a psychological test subject and image data based on an image having a pattern. As shown in FIG. 4, the image on the left (S310) is the original WZT image, and the image on the right (S320) is a composite image through the analysis unit.

[0053] Specifically, image data can be generated and output as new, image-enhanced data by synthesizing a background onto WZT images using a style transformation function for pre-training and application of the classification model. The image data can be synthesized from public domain photographs or images to the extent that the original image is not damaged.

[0054] Here, the style conversion function can composite the content image corresponding to the image data as a base and apply the style image corresponding to the image data.

[0055] Image classification performance varies depending on the trained images, and augmenting various images while preserving image features can be a method to prevent performance degradation.

[0056] GAN style transformation functions can be used to enhance images while maintaining the features of images drawn by dementia test subjects.

[0057] Due to the nature of images drawn by dementia test subjects, there is a significant amount of white space. When there is a lot of white space, there is a disadvantage in that additional time and resources are consumed to identify specific patterns for psychological testing, and accuracy is reduced.

[0058] Accordingly, to compensate for the aforementioned drawbacks, applying a style transformation function, which is one of the GAN methodologies, results in no distortion to the images created by the psychological test subjects; instead, by combining noise generated by GANs with the whitespace to form patterns, it has the effect of training the whitespace as noise.

[0059] Meanwhile, the present invention utilizes a stored and pre-trained artificial neural network model. The pre-trained artificial neural network model is a judgment model trained based on a plurality of images based on an artificial intelligence algorithm, and may be a model based on a neural network. The pre-trained artificial neural network model may be designed to simulate the structure of the human brain on a computer and may include a plurality of network nodes having weights that simulate neurons of a human neural network. The plurality of network nodes may each form a connection relationship to simulate the synaptic activity of neurons transmitting and receiving signals through synapses.

[0060] For example, it may include machine learning models, neural network models, or deep learning models that have evolved from neural network models. In deep learning models, multiple network nodes may be located at different depths (or layers) and exchange data according to convolutional connection relationships.

[0061] The trained artificial neural network model of the present invention is based on SRGAN (Super-resolution GAN) and can restore low-resolution images into high-resolution images.

[0062] Meanwhile, the table below shows the early-stage dementia data collected from 97 individuals by the Y Senior Welfare Division in City J from January to July 2024. As shown in the table below, the number of non-early-stage dementia data was also 97, and the data was analyzed using 82% (1,280 images) for training and 18% (272 images) for testing. A CNN technique was used to analyze the images while preserving the characteristics of the WZT drawings directly created by the elderly. Additionally, WZT images generated after training with computer vision and SRGAN were utilized for model training, as shown in Table 1.

[0063] [unit : %, piece]datasetClasstraining datasettestdatasetTotalratio (%)82%18%100%Drawing image piece1,2802721,552

[0064]

[0065] Meanwhile, Figures 5 and 6 are diagrams illustrating the learning and dementia determination procedures using CNN in an early dementia prediction device for the elderly according to an embodiment of the present invention.

[0066] The artificial neural network model of the present invention may be a Convolutional Neural Network (CNN) model trained based on dementia images. A CNN is a multilayer neural network with a special connection structure designed for speech processing, image processing, etc. Meanwhile, it goes without saying that the trained artificial neural network model is not limited to a CNN. For example, the training network model may be implemented as at least one Deep Neural Network (DNN) model among Recurrent Neural Networks (RNN), Long Short Term Memory Networks (LSTM), Gated Recurrent Units (GRU), or Generative Adversarial Networks (GAN).

[0067] CNNs consist of convolution layers, activation layers, pooling layers, and fully connected layers.

[0068] The activation functions used in the CNN layers are the ReLU function and the softmax function, and the optimizer can be Adam and the loss function can be the "sparse_categorical_crossentropy" function.

[0069] The epoch is performed up to 30 times, and the training is designed to terminate if it is determined that there is no performance improvement in validation accuracy for 3 or more times.

[0070] The result of the softmax function is interpreted as a probability value for "N" and "Y". For example, if the result is [0.2, 0.8], the probability of normal cognitive function is 20% and the probability of abnormal cognitive function (Y) is 80%, and the one with the higher probability can be selected as the final result.

[0071] In addition, the analysis unit (S400) of the present invention can predict whether a subject with early-stage dementia has dementia by using an artificial neural network model, and the artificial neural network model can utilize a Convolutional Neural Network (CNN) model based on a Soft Max function.

[0072] Specifically, the dementia prediction unit may be characterized by deriving a learning result based on a CNN model based on a softmax function, dividing the probability of normal cognitive function and the probability of abnormal cognitive function based on the softmax function to determine whether there is a cognitive impairment, selecting the probability of normal cognitive function and the probability of abnormal cognitive function as the final result, and determining that the cognitive function is normal when the probability of normal cognitive function is high, and determining that the cognitive function is abnormal when the probability of abnormal cognitive function is high.

[0073] Additionally, when the analysis unit (S400) derives a learning result, it can derive a learning result in which the cognitive function is classified as abnormal if the resulting number is negative and as normal if the number is positive, based on a softmax function, and the cognitive function is classified as normal if the number is negative.

[0074] Meanwhile, the output unit (130) according to one embodiment of the present invention outputs dementia prediction information regarding the degree of dementia of the examiner analyzed from the analysis unit (120).

[0075] The communication unit (140) of the early dementia prediction device of the present invention performs the function of enabling the dementia prediction device to communicate with an external device.

[0076] The control unit (150) of the early dementia prediction device of the present invention performs the function of controlling the flow of data between the data collection unit (110), the analysis unit (120), the output unit (130), and the communication unit (140). That is, the control unit (150) according to the present invention controls the flow of data from outside the early dementia prediction device or between each component of the dementia prediction device, thereby controlling the data collection unit (110), the analysis unit (120), the output unit (130), and the communication unit (140) to perform their respective unique functions.

[0077] According to one embodiment of the present invention, the data collection unit (110), analysis unit (120), output unit (130), communication unit (140), and control unit (150) may be program modules, at least some of which communicate with an external system. Such program modules may be included in an early dementia prediction device for the elderly utilizing artificial intelligence in the form of an operating system, an application program module, or other program modules, and may be physically stored in various known memory devices. Additionally, such program modules may be stored in a remote memory device capable of communicating with the early dementia prediction device for the elderly utilizing artificial intelligence. Meanwhile, such program modules encompass, but are not limited to, routines, subroutines, programs, objects, components, data structures, etc., that perform specific tasks or execute specific abstract data types as described below according to the present invention.

[0078] FIG. 7 is a diagram illustrating a method for predicting early dementia in the elderly using artificial intelligence according to an embodiment of the present invention.

[0079] First, to collect data for dementia testing, the examiner draws additional pictures on an image with printed stimulus symbols and collects the corresponding image data (S110). The images collected at this time are based on the Wartegg Zeichen Test (WZT), which is a projection test result using various types of images on eight stimulus images with a 4x4cm black border.

[0080] Next, machine learning is performed using an artificial neural network model stored in the image data, and the degree of dementia prediction is analyzed (S120). At this time, to analyze the degree of dementia prediction, any one of the Canny Edge Detection method, SRGAN (Super-Resolution Generative Adversarial Network), and CNN (Convolutional Neural Network) techniques may be used.

[0081] In addition, depending on the implementation, an examiner may first draw additional pictures on an image with printed stimulus symbols, and then apply CNN techniques to the image to collect image data. Machine learning may also be performed on the CNN-trained image using either the Canny Edge Detection method or SRGAN (Super-Resolution Generative Adversarial Network) to analyze the degree of dementia prediction.

[0082] Subsequently, dementia prediction information regarding the degree of dementia of the examiner analyzed by the analysis unit can be output (S130). The output unit refers to a terminal capable of transmitting and receiving wired and wirelessly, such as a PDA, PC, laptop computer, mobile phone, etc., that provides visual and voice information, and any one of all terminals capable of visualizing the dementia data predicted by the analysis unit may be adopted.

[0083] In this way, the present invention enables the prediction of whether a test subject has early-stage dementia by performing learning using various artificial neural network-based machine learning models on data based on images expressed by the Wartegg-Zeichen test (WZT).

[0084] In addition, unlike conventional methods where experts primarily determined the presence of dementia through paper surveys or cognitive tests, the present invention enables the determination of dementia using a simple treadmill model.

[0085] Therefore, it results in a reduction of the costs associated with using experts and avoids incurring significant expenses.

[0086] The device described above may be implemented as a hardware component, a software component, and / or a combination of a hardware component and a software component. For example, the device and components described in the embodiments may be implemented using one or more general-purpose or special-purpose computers, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing unit may execute an operating system (OS) and one or more software applications executed on the operating system. Additionally, the processing unit may access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing unit may be described as being used as a single unit, but those skilled in the art will understand that the processing unit may include multiple processing elements and / or multiple types of processing elements. For example, the processing unit may include multiple processors or one processor and one controller. In addition, other processing configurations, such as parallel processors, are also possible.

[0087] Software may include computer programs, code, instructions, or a combination of one or more of these, and may configure a processing unit to operate as desired or instruct the processing unit independently or collectively. Software and / or data may be embodied in any type of machine, component, physical device, virtual equipment, computer storage medium, or device so as to be interpreted by the processing unit or to provide instructions or data to the processing unit. Software may be distributed over networked computer systems and may be stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

[0088] The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., either alone or in combination. The program instructions recorded on the medium may be those specifically designed and configured for the embodiment, or they may be those known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc.

[0089] Although the embodiments have been described above with reference to limited examples and drawings, those skilled in the art can make various modifications and variations from the description above. For example, suitable results may be achieved even if the described techniques are performed in a different order than described, and / or if the components of the described system, structure, device, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents. Therefore, other implementations, other embodiments, and equivalents to the claims below are also within the scope of the claims.

Claims

1. An image data collection unit in which an examiner draws additional pictures on an image with printed stimulus symbols and collects image data; An analysis unit that performs machine learning with an artificial neural network model stored in the above image data and analyzes the degree of dementia prediction; Characterized by including an output unit that outputs dementia prediction information regarding the degree of dementia of the examiner analyzed from the analysis unit. Early dementia prediction device for the elderly using artificial intelligence.

2. In Paragraph 1, The image collected by the above data collection unit is characterized by being composed of the WZT (Wartegg Zeichen Test). Early dementia prediction device for the elderly using artificial intelligence.

3. In Paragraph 1, The analysis unit is characterized by being composed of a Canny Edge Detection method that divides the pixels of the image data into different intervals and verifies the boundaries of the image data. Early dementia prediction device for the elderly using artificial intelligence.

4. In Paragraph 1, The above analysis unit is characterized by using SRGAN (Super-Resolution Generative Adversarial Network) to convert the image of the above image data into a high-resolution image. Early dementia prediction device for the elderly using artificial intelligence.

5. In Paragraph 1, The analysis unit described above is characterized by learning the image data using the CNN (Convolutional Neural Network) technique. Early dementia prediction device for the elderly using artificial intelligence.

6. In Paragraph 5, The above CNN is characterized by being composed of CNN layers, which are a convolution layer, an activation layer, a pooling layer, and a fully connected layer. Early dementia prediction device for the elderly using artificial intelligence.

7. In Paragraph 6, The activation function used in the above CNN layer is characterized by using the ReLU (Rectified Linear Unit) function and the Softmax function, Early dementia prediction device for the elderly using artificial intelligence.

8. In Paragraph 1, The analysis unit derives a learning result based on a softmax function, derives a probability of normal cognitive function and a probability of abnormal cognitive function based on the softmax function regarding the presence of dementia, selects the probability of normal cognitive function and the probability of abnormal cognitive function as the final result, and is characterized by determining that cognitive function is normal when the probability of normal cognitive function is high, and determining that cognitive function is abnormal when the probability of abnormal cognitive function is high. Early dementia prediction device for the elderly using artificial intelligence.

9. In Paragraph 8, The analysis unit is characterized by deriving a learning result in which the above-described number is classified as cognitive function abnormal if it is negative and as cognitive function normal if it is positive, by applying a linear function to a layer that processes the above-described normal probability of cognitive function and the above-described abnormal probability of cognitive function. Early dementia prediction device for the elderly using artificial intelligence.

10. An image data collection step in which the examiner draws additional pictures on an image printed with stimulus symbols and collects image data; A dementia prediction analysis step that performs machine learning on image data collected in the above data collection step using an artificial neural network model stored therein and analyzes the degree of dementia prediction; Characterized by including an output step for outputting dementia prediction information regarding the degree of dementia of the examiner analyzed from the above dementia prediction analysis step. Method for predicting early dementia in the elderly using artificial intelligence.

11. In Paragraph 10, The image collected in the above data collection step is characterized by being composed of the WZT (Wartegg Zeichen Test). Method for predicting early dementia in the elderly using artificial intelligence.

12. In Paragraph 10, The above dementia prediction analysis step is characterized by being composed of a Canny Edge Detection method that divides the pixels of the image data into different intervals and verifies the boundaries of the image data. Method for predicting early dementia in the elderly using artificial intelligence.

13. In Paragraph 10, The above dementia prediction analysis step is characterized by using SRGAN (Super-Resolution Generative Adversarial Network) to convert the image of the image data into a high-resolution image. Method for predicting early dementia in the elderly using artificial intelligence.

14. In Paragraph 10, The above dementia prediction analysis step is characterized by learning the image data using a CNN (Convolutional Neural Network) technique. Method for predicting early dementia in the elderly using artificial intelligence.

15. An image data collection unit that collects image data by having an examiner draw additional pictures on an image with printed stimulus symbols and applying a CNN (Convolutional Neural Network) technique to the image; An analysis unit that performs machine learning on the above image data using either the Canny Edge Detection method or SRGAN (Super-Resolution Generative Adversarial Network) and analyzes the degree of dementia prediction; Characterized by including an output unit that outputs dementia prediction information regarding the degree of dementia of the examiner analyzed from the analysis unit. Early dementia prediction device for the elderly using artificial intelligence.

Images