CN111739636A - PPAT-based psychological intelligent analysis system - Google Patents

PPAT-based psychological intelligent analysis system Download PDF

Info

Publication number
CN111739636A
CN111739636A CN202010564323.9A CN202010564323A CN111739636A CN 111739636 A CN111739636 A CN 111739636A CN 202010564323 A CN202010564323 A CN 202010564323A CN 111739636 A CN111739636 A CN 111739636A
Authority
CN
China
Prior art keywords
ppat
image
dimension
algorithm
theme
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202010564323.9A
Other languages
Chinese (zh)
Inventor
韩易静
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhien Peixin Beijing Technology Co ltd
Original Assignee
Zhien Peixin Beijing Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Zhien Peixin Beijing Technology Co ltd filed Critical Zhien Peixin Beijing Technology Co ltd
Priority to CN202010564323.9A priority Critical patent/CN111739636A/en
Publication of CN111739636A publication Critical patent/CN111739636A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/56—Extraction of image or video features relating to colour
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Medical Informatics (AREA)
  • Biomedical Technology (AREA)
  • Software Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Public Health (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • General Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Databases & Information Systems (AREA)
  • Molecular Biology (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Multimedia (AREA)
  • Pathology (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

The invention discloses a PPAT-based mental intelligent analysis system, which relates to the technical field of image classification and comprises a drawing acquisition subsystem and a drawing mental intelligent analysis subsystem; the drawing acquisition subsystem is a drawing system for providing a tester with an image drawing function; the drawing acquisition subsystem is used for drawing the PPAT theme image; and the painting mental intelligent analysis subsystem is used for processing the acquired PPAT theme image according to the image feature extraction algorithm and the machine learning algorithm and determining a depression tendency judgment result corresponding to the PPAT theme image. The PPAT psychological measurement tool can automatically analyze the PPAT topic image drawn by a tester, reduce the dependence on the priori psychological knowledge of professional therapists, and improve the popularity and the usability of the PPAT psychological measurement tool.

Description

PPAT-based psychological intelligent analysis system
Technical Field
The invention relates to the technical field of image classification, in particular to a PPAT-based mental intelligent analysis system.
Background
PPAT (personnickingand applefrom tree) is a psychometric tool designed in conjunction with a painting projection test, which proposes a task of drawing a PPAT subject image to a tester, and after the tester draws the PPAT subject image, the PPAT subject image drawn by the tester is scored in multiple dimensions by a therapist, thereby identifying depression, manic-depressive illness, schizophrenia, senile dementia, and other cognitive disorders according to the scores. At present, the application of identifying the depression tendency according to a PPAT psychometric tool is less in China, professional therapists are required to score PPAT theme images drawn by testers, whether the testers have the depression tendency or not is comprehensively judged according to scores, the requirements on psychological knowledge and experience of the testers are high, and the psychometric technology is difficult to apply and popularize on a large scale.
Disclosure of Invention
The invention aims to provide a PPAT-based psychological intelligent analysis system, which combines a PPAT psychological measurement tool with an intelligent analysis technology, automatically analyzes a PPAT subject image drawn by a tester, reduces the dependence on the priori psychological knowledge of a professional therapist, and improves the popularity and the usability of the PPAT psychological measurement tool.
In order to achieve the purpose, the invention provides the following scheme:
a PPAT-based mental intelligent analysis system comprises a drawing acquisition subsystem and a drawing mental intelligent analysis subsystem; the drawing acquisition subsystem is a drawing system for providing a tester with an image drawing function; an image feature extraction algorithm and a machine learning algorithm are built in the painting mental intelligent analysis subsystem;
the drawing acquisition subsystem is used for drawing a PPAT theme image and sending the PPAT theme image to the drawing mental intelligent analysis subsystem;
the drawing mental intelligent analysis subsystem is used for processing the PPAT theme image according to the image feature extraction algorithm and the machine learning algorithm and determining a depression tendency judgment result corresponding to the PPAT theme image.
Optionally, the drawing collection subsystem includes a user-side drawing device and an image collection processor;
the user side drawing equipment is used for drawing a PPAT theme image;
the image acquisition processor is used for acquiring the PPAT theme image and the basic data corresponding to the PPAT theme image and sending the PPAT theme image and the basic data corresponding to the PPAT theme image to the painting mental intelligence analysis subsystem.
Optionally, the user drawing device is a tablet or tablet device.
Optionally, the image feature extraction algorithm includes a convolutional neural network algorithm and a conventional image processing algorithm; the machine learning algorithm comprises a support vector machine algorithm, a random forest algorithm and a multilayer perceptron algorithm.
Optionally, the painting mental intelligence analysis subsystem comprises a plurality of dimension graders and a depression tendency classifier connected with each dimension grader; wherein the number of the dimension graders is determined according to the dimension rule of the PPAT psychometric measuring tool.
Optionally, the dimension scorer is configured to perform image feature extraction processing on the PPAT subject image, score the PPAT subject image according to the extracted image feature, and send the score of the PPAT subject image to the depression tendency classifier;
the depression tendency classifier is used for determining a depression tendency judgment result corresponding to the PPAT theme image according to the score of the PPAT theme image.
Optionally, the number of the dimension scorers is 7, and the dimension scorers are respectively a color application dimension scorer, a color proper dimension scorer, an energy allocation dimension scorer, a space use dimension scorer, a true degree dimension scorer, a detail description dimension scorer and a human body integrity dimension scorer.
Optionally, the depression tendency classifier is a classifier which is classified two and is constructed by using a support vector machine algorithm, a random forest algorithm or a multi-layer perceptron algorithm.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention provides a PPAT-based mental intelligent analysis system which can automatically analyze a PPAT theme image drawn by a drawing acquisition subsystem and directly calculate whether a tester has a depression tendency according to an analysis result. Therefore, the PPAT psychological measurement tool can reduce the dependence on the priori psychological knowledge of professional therapists and improve the popularity and the usability of the PPAT psychological measurement tool.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
FIG. 1 is a block diagram of a PPAT-based mental intelligence analysis system according to the present invention;
FIG. 2 is a block diagram of the process of the intelligent analysis subsystem for painting psychology according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention aims to provide a PPAT-based psychological intelligent analysis system, which combines a PPAT psychological measurement tool with an intelligent analysis technology, automatically analyzes a PPAT subject image drawn by a tester, reduces the dependence on the priori psychological knowledge of a professional therapist, and improves the popularity and the usability of the PPAT psychological measurement tool.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
PPAT: PersonPickingand ApplefromaTree, a psychometric tool designed in conjunction with a painting projection test, which presents the tester with the task of drawing a PPAT subject image, and after the tester draws the PPAT subject image, the tester draws the PPAT subject image by a therapist by scoring the PPAT subject image drawn by the tester in multiple dimensions, thereby identifying depression, manic-depressive illness, schizophrenia, senile dementia, and other cognitive disorders according to the scores.
The intelligent analysis technology comprises the following steps: by means of an artificial intelligence technology, image features can be automatically extracted from image data, so that a complex image classification and identification task is solved, and the method has many successful applications in the field of images. The intelligent analysis technology can be trained according to historical image data and big data, automatically learns, and can complete image classification and recognition tasks in a testing link without or with less need of human professional knowledge.
Obviously, the intelligent analysis technology can automatically analyze and classify images, so that the problems that a large amount of professional knowledge is needed in psychological measurement and large-scale popularization is difficult are solved; therefore, the PPAT psychometric measuring tool and the intelligent analysis technology are combined, and the method has wide application prospects.
As shown in fig. 1, the present embodiment provides a PPAT-based mental intelligence analysis system, which includes a drawing acquisition subsystem and a drawing mental intelligence analysis subsystem. The drawing acquisition subsystem is a drawing system for providing a tester with an image drawing function; an image feature extraction algorithm and a machine learning algorithm are built in the painting mental intelligent analysis subsystem.
The drawing acquisition subsystem is used for drawing the PPAT theme image and sending the PPAT theme image to the drawing mental intelligent analysis subsystem.
The drawing mental intelligent analysis subsystem is used for processing the PPAT theme image according to the image feature extraction algorithm and the machine learning algorithm and determining a depression tendency judgment result corresponding to the PPAT theme image, so that whether a tester drawing the PPAT theme image has a depression tendency or not is determined.
In this embodiment, the drawing collection subsystem includes a user-side drawing device and an image collection processor; the painting mental intelligence analysis subsystem comprises a plurality of dimension graders and a depression tendency classifier connected with each dimension grader. Wherein the number of the dimension graders is determined according to the dimension rule of the PPAT psychometric measuring tool.
A tester draws a PPAT theme image by operating the drawing acquisition subsystem, and then analyzes the PPAT theme image drawn by the drawing acquisition subsystem by the drawing mental intelligent analysis subsystem to obtain whether the tester drawing the PPAT theme image has a depression tendency or not, so that the problems that a large amount of professional knowledge is needed in psychological measurement and large-scale popularization is difficult are solved.
Preferably, the drawing collection subsystem provided by this embodiment uses the user-side drawing device to provide the tester with the function of drawing the PPAT theme image, and uses the image collection processor to collect the PPAT theme image drawn by the tester.
The user side drawing device is a digital board or a flat panel device and is used for drawing the PPAT theme image. The image acquisition processor is used for acquiring the PPAT theme image and the basic data corresponding to the PPAT theme image and sending the PPAT theme image and the basic data corresponding to the PPAT theme image to the painting mental intelligence analysis subsystem.
Preferably, a specific flow of the painting mental intelligence analysis subsystem provided in this embodiment is shown in fig. 2, and specifically includes: the method comprises the steps of inputting a PPAT theme image acquired by a drawing acquisition subsystem, extracting image features from the PPAT theme image by using a convolutional neural network algorithm or a traditional image processing algorithm, classifying the image features by using machine learning methods such as a neural network algorithm and a support vector machine algorithm to obtain the score of the PPAT theme image in each dimension, classifying the PPAT theme image by using a support vector machine algorithm, a random forest algorithm or a multi-layer perceptron algorithm according to the scores of the PPAT theme image in multiple dimensions, and determining a depression tendency judgment result corresponding to the PPAT theme image, namely outputting the depression tendency judgment result.
The dimension scorer provided by the embodiment is used for extracting the image features of the PPAT theme image and scoring the PPAT theme image according to the image features.
The PPAT psychometric tool provided by this embodiment includes 7 dimensions including color application, that is, in this embodiment, 7 dimension scorers are provided, and the 7 dimension scorers are adopted to score the input PPAT subject image.
The specific process of the dimension scorer is as follows: the input is a PPAT theme image, each dimension scorer extracts corresponding image features from the PPAT theme image in the dimension of the corresponding dimension, scores the PPAT theme image in the dimension to obtain the score of the PPAT theme image in the dimension, and then sends the 7 scores to the depression tendency classifier.
The rules and their algorithmic flows for each dimension scorer are described in detail below.
First, color application dimension scorer
# 1-prominenceofcolor (application of color)
The color application score is measured by the number of color types used in the PPAT theme image and the color coverage area, and the specific rule is as follows.
If no drawing is performed, i.e. the number of color categories is 0, the rating of the PPAT subject image in this dimension is 0.
If the object is outlined or drawn using color and the interior of the object is not filled with color, the PPAT theme image will have a rating of 1 in this dimension.
If most objects are outlined with color and only one object is filled with color, the PPAT subject image has a rating of 2 in this dimension.
If two or more, but not all, objects are filled with color, the PPAT subject image has a rating of 3 in this dimension.
If an object is outlined using color and filled in, the PPAT subject image has a rating of 4 in this dimension.
If not only the object is outlined and filled in with color, but also the object space such as the background is filled in, the rating of the PPAT theme image in this dimension is 5.
The algorithm flow of the color application dimension scorer is as follows: the input is a PPAT theme image, and the output is the score of the PPAT theme image in the color application dimension, and the specific algorithm steps are as follows:
1. and processing the PPAT theme image by using an image processing method, and calculating the number of color types.
2. The PPAT theme image is processed by using an image processing method, and the number of pixel points of each color is calculated.
3. And classifying the PPAT theme image by using a support vector machine algorithm according to the number of the color types and the number of the pixel points of each color to obtain the score of the PPAT theme image in the color application dimension, and outputting the score.
Second, color appropriate dimension scorer
# 2-ColorFit (color appropriate)
The appropriate color score is measured by whether the color type used by the painting object in the PPAT subject image is the appropriate color type of the object, and the specific rule is as follows.
If drawing is not performed or different colors cannot be distinguished from the PPAT subject image, the grade of the PPAT subject image in the dimension is 0.
If the PPAT subject image is drawn using only one color, and the color is among six colors of cyan, pink, magenta, yellow, and violet, the PPAT subject image has a rating of 1 in this dimension.
If the PPAT subject image is drawn using only one color, and the colors are red, green, dark green, brown, and black, the rating of the PPAT subject image in this dimension is 2.
If the colors of a part of objects but not all objects in the PPAT theme image are properly used, that is, the colors used by the part of objects are in the proper color set of the part of objects, the grade of the PPAT theme image in the dimension is 3.
If most colors in the PPAT subject image are well utilized, the grade of the PPAT subject image in the dimension is 4.
If all colors in the PPAT subject image are well applied, the rating of the PPAT subject image in this dimension is 5.
The algorithm flow of the color proper dimension scorer is as follows: the input is a PPAT theme image, and the output is the score of the PPAT theme image in the appropriate dimension of color, and the specific algorithm steps are as follows:
1. and processing the PPAT theme image by using an image processing method, and calculating the number of color types.
2. Determining an output score according to the number of the color types, specifically:
if the number of color classes is equal to 0, the output score is 0.
If the number of color categories is equal to 1 and the color categories are in the color set of cyan, pink, magenta, yellow, violet, the output score is 1.
If the number of color categories is equal to 1 and the color categories are in the color set of red, green, dark green, brown, black, the output score is 2.
If the number of color types is greater than 1, go to step 3.
3. Processing the PPAT theme image by using an image segmentation algorithm (segmenting a crown, a trunk, an apple and a human body in the PPAT theme image), and then judging whether the color of the part is suitable according to a segmentation result. In the PPAT theory, the color of the crown is considered to be proper when the color is red, green and dark green; the color of the trunk is proper when the trunk is brown and black; the color of apple is proper when it is red, yellow, green and dark green; the color of human body is proper when it is black, brown, yellow, orange, red. If the tree crown, the tree trunk, the apple and the human body are all properly filled, the output score is 5; if there are less than 2 parts correctly filled, the output score is 3; the other output scores were 4.
The image segmentation algorithm specifically used is described as follows:
dividing a PPAT theme image into different semantic regions, wherein the PPAT theme image mainly comprises three stages, wherein the first two stages are used for training two FastR-CNN (fast-regional convolutional neural network) models independently, one is used for detecting the whole human body and the whole tree, and the other is used for detecting each part in the image, such as a crown, a trunk, an apple and the like, so as to obtain the general position of each part; and in the third stage, the trained FastR-CNN model is used for segmenting each detected part, and finally, the segmentation result of each part is restored to the original image coordinate to obtain the final segmentation result.
After the segmentation result is obtained, color recognition is carried out on each part by using an image color histogram algorithm in the traditional image processing algorithm, and the color with the largest area is regarded as the color of the part. The colors are then compared to determine if the part color is in the proper color set for the part.
Third, effort adjustment dimension scorer
# 3-ImpliedEnergy (energy adjustment)
The stamina rating is measured by the tester's effort in the PPAT subject image, and the specific rules are as follows.
If no drawing is performed, the rating of the PPAT subject image in this dimension is 0.
If the PPAT subject image only completes the task requirement, the grade of the PPAT subject image in the dimension is 1.
If an additional effort is put into the PPAT theme image, the rating of the PPAT theme image in this dimension is 2.
If the effort expended in the PPAT subject image reaches an average level, the PPAT subject image has a rating of 3 in this dimension.
If much effort is put into the PPAT theme image, the PPAT theme image has a rating of 4 in this dimension.
If too much effort is put into the PPAT theme image, the PPAT theme image has a rating of 5 in this dimension.
The algorithm flow of the energy allocation dimension scorer is as follows: the input is a PPAT theme image, and the output is the score of the PPAT theme image in the stamina deployment dimension, and the specific algorithm steps are as follows:
1. image features in the PPAT subject image are extracted using the ResNet50 network model in the convolutional neural network.
2. And classifying the extracted image features in 6 score levels by using two fully connected layers, and outputting scores of the PPAT subject images in the deployment dimension.
The energy adjustment dimension scorer is obtained by training according to historical data or big data.
Fourth, the dimension of space usage is scored
# 4-Space (Space usage)
The score of the space usage is measured by the size proportion of the space occupied by the drawing area in the PPAT subject image in the whole image, and the specific rule is as follows.
If no drawing is performed, the grading level of the PPAT theme image in the dimension is 0;
if the proportion of the drawing area to the whole image area is lower than 25%, the grade of the PPAT theme image in the dimension is 1.
If the proportion of the drawing area to the whole image area is 25%, the rating of the PPAT subject image in the dimension is 2.
If the proportion of the drawing area to the whole image area is 50%, the rating of the PPAT subject image in the dimension is 3.
If the proportion of the drawing area to the whole image area is 75%, the rating of the PPAT subject image in the dimension is 4.
If the proportion of the drawing area to the whole image area is 100%, the rating of the PPAT subject image in the dimension is 5.
The algorithm flow of the space use dimension scorer is as follows: the input is a PPAT theme image, and the output is the score of the PPAT theme image in the space use dimension, and the specific algorithm steps are as follows:
1. and processing the PPAT theme image by using an image processing method, and calculating the area occupied by the circumscribed frame of the drawing content.
2. And calculating the ratio of the area occupied by the external frame to the total area of the canvas to obtain the drawing area ratio.
3. Determining an output score according to the drawing area proportion, specifically:
if the drawing area ratio is equal to 0, the output score is 0.
If the drawing area ratio is more than 0 and less than 25%, the output score is 1.
If the drawing area ratio is more than 25% and less than 50%, the output score is 2.
If the drawing area ratio is more than 50% and less than 75%, the output score is 3.
If the drawing area ratio is more than 75 and less than 100%, the output score is 4.
If the drawing area ratio is equal to 100%, the output score is 5.
Fifthly, true degree dimension scorer
# 5-Realism (true)
The score of the truth degree is measured by the image truth degree in the PPAT subject image, and the specific rule is as follows.
If only a large number of lines or shapes are visible in the PPAT subject image and any element cannot be recognized, the rating of the PPAT subject image in this dimension is 0.
If some elements in the PPAT subject image can be recognized as a human body, an apple or a tree, but the recognition confidence is low, the grade of the PPAT subject image in the dimension is 1.
If the object in the PPAT subject image is recognizable but the drawing is simple (if the trunk is represented by only one line), the grade of the PPAT subject image in the dimension is 2.
If the objects in the PPAT subject image are recognizable and there are slightly complex objects (e.g., a tree with a trunk, branches, leaves), the PPAT subject image has a rating of 3 in this dimension.
If the objects in the PPAT subject image are all depicted relatively truly, the rating of the PPAT subject image in this dimension is 4.
If the object in the PPAT theme image is very realistic and the element is close to 3 dimensions, the rating of the PPAT theme image in this dimension is 5.
The algorithm flow of the true degree dimension scorer is as follows: the input is a PPAT theme image, and the output is the score of the PPAT theme image in the dimension of the true degree, and the specific algorithm steps are as follows:
1. image features in the PPAT subject image are extracted using the ResNet50 network model in the convolutional neural network.
2. And classifying the extracted image features in 6 score grades by using two fully-connected layers, and outputting scores of the PPAT subject images in a true degree dimension.
The truth dimension scorer is obtained by training according to historical data or big data.
Sixth, the detail description dimension scorer
# 6-detailsofObjectsandEnvironment (detailed description)
The score of the detail description is measured by the degree of the detail description of the image in the PPAT subject image, and the specific rule is as follows.
If an element in the PPAT subject image cannot be identified, the rating of the PPAT subject image in this dimension is 0.
If there is an object (human body, tree or apple) required by the task in the PPAT subject image, but there is no other element, the grade of the PPAT subject image in this dimension is 1.
If there is an object (human body, tree or apple) required by the task in the PPAT subject image and there is a horizon or some grass, the rating of the PPAT subject image in this dimension is 2.
If the PPAT subject image has objects (human body, tree or apple) required by the task and has horizon and other details (such as flowers, sun and the like), the grade of the PPAT subject image in the dimension is 3.
If there is an object (human body, tree or apple) in the PPAT subject image that is required by the task and there is another and more than one detail, the rating of the PPAT subject image in this dimension is 4.
If there is an object (human body, tree or apple) in the PPAT subject image that is required by the task and there is a great deal of other detail, the rating of the PPAT subject image in this dimension is 5.
The algorithm flow of the detail description dimension scorer is as follows: the input is a PPAT theme image, and the output is the score of the PPAT theme image in the detail description dimension, and the specific algorithm steps are as follows:
1. image features in the PPAT subject image are extracted using a ResNet network model in a convolutional neural network.
2. And classifying the extracted image features in 6 score levels by using two fully connected layers, and outputting scores of the PPAT topic images in the detail description dimension.
The detail description dimension scorer is obtained by training according to historical data or big data.
Seventh, human body integrity dimension scorer
# 7-Person (human body)
The score of the human body is measured by the human body truth degree, namely the human body integrity degree, in the PPAT theme image, and the specific rule is as follows.
If a human body cannot be identified in the PPAT subject image, the grade of the PPAT subject image in the dimension is 0.
If some elements in the PPAT theme image can be recognized as human bodies, but the recognition confidence is low, the grade of the PPAT theme image in the dimension is 1.
If the identified human body in the PPAT subject image is only a part of the body, not a complete human body, the grade of the PPAT subject image in the dimension is 2.
If the identified human body in the PPAT subject image is a matchmaker, who has only one circle as the head, the PPAT subject image has a rating of 3 in this dimension.
If the identified human body in the PPAT subject image is a matchmaker, who has facial features or a torso but lacks some body parts, the PPAT subject image has a rating of 4 in this dimension.
If the human body recognized in the PPAT subject image depicts various parts of the body, the grade of the PPAT subject image in the dimension is 5.
The algorithm flow of the human body integrity dimension scorer is as follows: the input is a PPAT theme image, and the output is the score of the PPAT theme image in the human body integrity dimension, and the specific algorithm steps are as follows:
1. and (3) segmenting the human body part in the PPAT theme image by using an image segmentation algorithm, and then processing the segmentation result again to only keep the human body part. Wherein the segmentation algorithm is an image segmentation algorithm used by the color-appropriate dimension scorer.
2. And (3) carrying out image feature extraction and classification on the data obtained in the step 1 by using a ResNet50 model, and outputting the score of the PPAT subject image in the human body integrity dimension.
The human body integrity dimension scorer is obtained by training according to historical data or big data.
Through the 7-dimension scorer, scores of the PPAT theme image in 7 dimensions can be obtained, and then the 7-dimension scores are input into a depression tendency classifier for depression tendency classification.
And the depression tendency classifier is used for determining a depression tendency judgment result corresponding to the PPAT theme image according to the score of the PPAT theme image. Wherein, the depression tendency classifier is a classifier which is classified by two and is constructed by using a support vector machine algorithm, a random forest algorithm or a multilayer perceptron algorithm.
The depression tendency classifier provided by the embodiment is obtained by training a binary classifier by adopting a support vector machine algorithm, a random forest algorithm or a multilayer perceptron algorithm according to historical data or big data, and the specific training process comprises the following steps: the input data of the training samples are scores of the PPAT theme images sent by the dimensionality scorers, the output data of the training samples are depression tendency judgment results corresponding to the PPAT theme images, and the two classification classifiers are trained to obtain the required depression tendency classifier by adopting a support vector machine algorithm, a random forest algorithm or a multi-layer perceptron algorithm according to the input data and the output data in the training samples. Wherein the depression tendency classifier comprises 2 categories of depression tendency and depression-free tendency.
Through the design, the dimension scorer and the depression tendency classifier can continuously train the scoring capacity and the classifying capacity, and as long as the training samples are sufficient, the classifying accuracy of the whole painting psychological intelligent analysis subsystem can be infinitely close to the real psychological state of a tester, so that whether the tester has depression tendency is determined.
The existing PPAT psychometric technology needs a professional therapist to analyze the PPAT theme image drawn by a tester, and is difficult to be widely applied. The PPAT-based psychological intelligent analysis system provided by the invention can automatically and intelligently analyze the PPAT theme image drawn by a tester, reduce the requirement on professional knowledge of professional therapists, and improve the usability of the PPAT psychological measurement tool, thereby improving the popularity of the PPAT psychological measurement tool.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.

Claims (8)

1. A PPAT-based mental intelligent analysis system is characterized by comprising a drawing acquisition subsystem and a drawing mental intelligent analysis subsystem; the drawing acquisition subsystem is a drawing system for providing a tester with an image drawing function; an image feature extraction algorithm and a machine learning algorithm are built in the painting mental intelligent analysis subsystem;
the drawing acquisition subsystem is used for drawing a PPAT theme image and sending the PPAT theme image to the drawing mental intelligent analysis subsystem;
the drawing mental intelligent analysis subsystem is used for processing the PPAT theme image according to the image feature extraction algorithm and the machine learning algorithm and determining a depression tendency judgment result corresponding to the PPAT theme image.
2. The PPAT-based mental intelligent analysis system of claim 1, wherein the drawing acquisition subsystem comprises a client drawing device and an image acquisition processor;
the user side drawing equipment is used for drawing a PPAT theme image;
the image acquisition processor is used for acquiring the PPAT theme image and the basic data corresponding to the PPAT theme image and sending the PPAT theme image and the basic data corresponding to the PPAT theme image to the painting mental intelligence analysis subsystem.
3. The PPAT-based mental intelligence analysis system of claim 2, wherein the user-side drawing device is a tablet or tablet device.
4. The PPAT-based mental intelligence analysis system of claim 1, wherein the image feature extraction algorithm comprises a convolutional neural network algorithm and a conventional image processing algorithm; the machine learning algorithm comprises a support vector machine algorithm, a random forest algorithm and a multilayer perceptron algorithm.
5. The PPAT-based mental intelligence analysis system of claim 1, wherein the pictorial mental intelligence analysis subsystem comprises a plurality of dimension scorers and a depression tendency classifier connected to each of the dimension scorers; wherein the number of the dimension graders is determined according to the dimension rule of the PPAT psychometric measuring tool.
6. The PPAT-based mental intelligence analysis system according to claim 5, wherein the dimension scorer is configured to perform image feature extraction processing on the PPAT subject image, score the PPAT subject image according to the extracted image feature, and then send the score of the PPAT subject image to the depression tendency classifier;
the depression tendency classifier is used for determining a depression tendency judgment result corresponding to the PPAT theme image according to the score of the PPAT theme image.
7. The PPAT-based mental intelligence analysis system of claim 5, wherein the number of the dimension scorers is 7, and the dimension scorers are a color application dimension scorer, a color proper dimension scorer, an energy allocation dimension scorer, a space use dimension scorer, a true degree dimension scorer, a detail description dimension scorer and a human body integrity dimension scorer.
8. The PPAT-based mental intelligence analysis system of claim 5, wherein the depression tendency classifier is a two-class classifier constructed using a support vector machine algorithm, a random forest algorithm, or a multi-layered perceptron algorithm.
CN202010564323.9A 2020-06-19 2020-06-19 PPAT-based psychological intelligent analysis system Pending CN111739636A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010564323.9A CN111739636A (en) 2020-06-19 2020-06-19 PPAT-based psychological intelligent analysis system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010564323.9A CN111739636A (en) 2020-06-19 2020-06-19 PPAT-based psychological intelligent analysis system

Publications (1)

Publication Number Publication Date
CN111739636A true CN111739636A (en) 2020-10-02

Family

ID=72650190

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010564323.9A Pending CN111739636A (en) 2020-06-19 2020-06-19 PPAT-based psychological intelligent analysis system

Country Status (1)

Country Link
CN (1) CN111739636A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113707275A (en) * 2021-08-27 2021-11-26 郑州铁路职业技术学院 Mental health estimation method and system based on big data analysis
CN114550918A (en) * 2022-02-23 2022-05-27 中国科学院心理研究所 Mental disorder evaluation method and system based on drawing characteristic data

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104715157A (en) * 2015-03-25 2015-06-17 成都信息工程学院 Cognition impairment evaluating system and method based on clock drawing test
CN106447042A (en) * 2016-08-31 2017-02-22 广州瑞基信息科技有限公司 Psychoanalysis method and apparatus based on drawing projection
CN108230427A (en) * 2018-01-19 2018-06-29 京东方科技集团股份有限公司 A kind of intelligence is drawn a picture equipment, picture analysis system and picture processing method
CN110211668A (en) * 2019-05-31 2019-09-06 邵阳学院 A kind of psychoanalysis method and device based on drawing psychology
CN110931129A (en) * 2019-12-10 2020-03-27 上海市精神卫生中心（上海市心理咨询培训中心） Painting and drawing computer analysis method for evaluating schizophrenia mental state
CN111091910A (en) * 2019-12-17 2020-05-01 中国科学院自动化研究所 Intelligent evaluation system based on drawing clock test

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104715157A (en) * 2015-03-25 2015-06-17 成都信息工程学院 Cognition impairment evaluating system and method based on clock drawing test
CN106447042A (en) * 2016-08-31 2017-02-22 广州瑞基信息科技有限公司 Psychoanalysis method and apparatus based on drawing projection
CN108230427A (en) * 2018-01-19 2018-06-29 京东方科技集团股份有限公司 A kind of intelligence is drawn a picture equipment, picture analysis system and picture processing method
CN110211668A (en) * 2019-05-31 2019-09-06 邵阳学院 A kind of psychoanalysis method and device based on drawing psychology
CN110931129A (en) * 2019-12-10 2020-03-27 上海市精神卫生中心（上海市心理咨询培训中心） Painting and drawing computer analysis method for evaluating schizophrenia mental state
CN111091910A (en) * 2019-12-17 2020-05-01 中国科学院自动化研究所 Intelligent evaluation system based on drawing clock test

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
黄继生: ""基于卷积神经网络的房树人绘画图像分类研究"", 《中国优秀硕士学位论文全文数据库哲学与人文科学辑》, pages 4 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113707275A (en) * 2021-08-27 2021-11-26 郑州铁路职业技术学院 Mental health estimation method and system based on big data analysis
CN113707275B (en) * 2021-08-27 2023-06-23 郑州铁路职业技术学院 Mental health estimation method and system based on big data analysis
CN114550918A (en) * 2022-02-23 2022-05-27 中国科学院心理研究所 Mental disorder evaluation method and system based on drawing characteristic data

Similar Documents

Publication Publication Date Title
CN104992142B (en) A kind of pedestrian recognition method being combined based on deep learning and attribute study
CN106611169B (en) A kind of dangerous driving behavior real-time detection method based on deep learning
CN106650806B (en) A kind of cooperating type depth net model methodology for pedestrian detection
CN106897738B (en) A kind of pedestrian detection method based on semi-supervised learning
CN105574550B (en) A kind of vehicle identification method and device
CN105608446B (en) A kind of detection method and device of video flowing anomalous event
CN105046277B (en) Robust mechanism study method of the feature significance in image quality evaluation
CN106650786A (en) Image recognition method based on multi-column convolutional neural network fuzzy evaluation
CN109063728A (en) A kind of fire image deep learning mode identification method
CN107977671A (en) A kind of tongue picture sorting technique based on multitask convolutional neural networks
CN107103298A (en) Chin-up number system and method for counting based on image procossing
CN108491077A (en) A kind of surface electromyogram signal gesture identification method for convolutional neural networks of being divided and ruled based on multithread
CN108229458A (en) A kind of intelligent flame recognition methods based on motion detection and multi-feature extraction
CN104166548B (en) Deep learning method based on Mental imagery eeg data
CN109345770A (en) A kind of child leaves in-vehicle alarm system and child leaves interior alarm method
CN105608432A (en) Instantaneous myoelectricity image based gesture identification method
CN104778466B (en) A kind of image attention method for detecting area for combining a variety of context cues
CN102034107B (en) Unhealthy image differentiating method based on robust visual attention feature and sparse representation
CN111739636A (en) PPAT-based psychological intelligent analysis system
CN109671274A (en) A kind of highway risk automatic evaluation method based on latent structure and fusion
CN107862692A (en) A kind of ribbon mark of break defect inspection method based on convolutional neural networks
CN103971106A (en) Multi-view human facial image gender identification method and device
CN106874929A (en) A kind of pearl sorting technique based on deep learning
Khatun et al. Fruits classification using convolutional neural network
CN114187664B (en) Rope skipping counting system based on artificial intelligence

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination