automated human-level diagnosis of dysgraphia using a consumer tablet - digital pen tablet-ITATOUC

automated human-level diagnosis of dysgraphia using a consumer tablet - digital pen tablet

by:ITATOUCH     2020-05-02
automated human-level diagnosis of dysgraphia using a consumer tablet  -  digital pen tablet
Children's academic and behavioral progress is related to the timely development of reading and writing skills.
Difficulties in writing are usually in coordination with dyslexia and development (dyspraxia)
Or impairment of attention, which is the nervous system.
Developmental disorders
Difficulties in writing can seriously harm children's daily life and require therapeutic care.
Therefore, it is very important to find handwriting difficulties early in Pediatrics.
Since the beginning of the 20 th century, many handwriting scales have been developed to assess the quality of handwriting.
However, these tests often involve experts investigating visual sentences written on paper on a topic, so they are subjective, expensive, and of a poor size.
In addition, they ignore potential important features of motor control, such as writing dynamics, pen pressure, or pen tilt.
However, with the increasing popularity of digital tablets, the ability to measure these neglected features is now likely to be available on a large scale and at a very low cost.
In this work, we have developed a diagnostic tool that requires only ordinary tablets.
To this end, we modeled the data of 298 children, including 56 children with writing difficulties.
The children tested BHK on a paper-covered digital tablet.
We extracted 53 handwritten features to describe all aspects of handwriting and used a random forest classifier to diagnose writing difficulties.
Our approach is 96.
6% sensibility 99. 2% specificity.
Considering the internalJudges and International
In BHK testing, our technology has considerable accuracy for experts and can be deployed directly as diagnostic tools.
Despite the widespread use of laptops and tablets in schools, handwriting remains an important skill to acquire during the education of children, as it is the foundation of core educational activities such as taking notes, writing, and self-expression.
Handwriting is a complex task because it involves attention, perception, language, and fine motor skills.
This is why, even in children with normal development, the time spent learning calligraphy between the ages of 5 has crossed 10 years (preschool)and 15.
During this period, handwriting was initially developed at a qualitative level (legibility)
Then in quantity (speed).
Even with the right training, 5 to 34% of children will not master handwriting.
As their cognitive needs for school work continue to rise during school learning, these children will soon face more common difficulties.
When they are in trouble with automated handwriting, they cannot handle tasks such as grammar, spelling, and composition at the same time.
This leads to increased fatigue, decreased cognitive and self-abilityesteem.
Therefore, it is essential to identify and remedy any handwritten difficulties as early as possible.
Deuel presents the most mature and popular classification of writing difficulties.
Three points for supplies
The type of disease :(1)
Dyslexia occurs when the text written spontaneously is illegible and a copy of the written text is relatively preserved; (2)
Space writing is difficult. this is due to the defects in the understanding of space. its features are illegible writing, whether it is spontaneous or copied, and the writing speed remains normal; and (3)
When both spontaneously written and copied texts may be illegible, difficulty in motor writing that reflects motor impairment occurs.
In this type of writing difficulty, the writing speed and drawing are not normal.
Many quantitative tests have been proposed to evaluate calligraphy.
Most quantitative methods evaluate handwriting based on several predefined specific criteria.
The experts then rate these criteria and willscores.
Some tests using this principle have been developed for different letters.
Most tests that assess difficulty in writing are based on replication tasks (see Table )
, Which means that dyslexia cannot be detected using these tests.
In the table, we summarize the different tests that are widely used to diagnose writing difficulties.
As shown in the figure, these tests are heterogeneous because they are specifically designed to assess the quality of writing for a specific alphabet or age range.
In addition, we can see that these tests are handwritten based on different writing tasks (
Core tasks in table)
, Which may mean a high degree of variability in the results.
Finally, an important part of the entire handwriting process is not taken into account.
Only handwritten end products are used for analysis, regardless of the dynamics of handwriting, tilt, and pressure in most cases.
A major drawback of these tests is that the scores of several parameters depend on human judgment, which makes the test more subjective.
In addition, it is time to score the BHK exam.
Because it can take 15 minutes to score.
In addition, since the expert in charge of scoring can only access the final static image of the child's handwriting, some handwriting aspects such as handwriting dynamics with very rich information, the pressure between the pen and the tablet and the tilt of the pen are still hidden, so it is not used in the diagnosis.
Similarly, posture and grasp style are difficult to evaluate and must be done on site by an expert evaluator.
Finally, standardize the text used in the test (
The content of the text is always the same).
Therefore, tests cannot be conducted during the ecological writing conference (e. g.
, Text written by children every day during school meetings).
The rapid development of digital tablets over the past decade has enabled us to solve some of these problems in part.
This makes it possible to evaluate not only the handwritten final product (
Static image)
The same is true of its dynamics.
To better understand the barriers to writing, several studies have adopted these new technologies.
Pagliarini and others.
Before hand writing automatically, collect data on handwriting ability using a tablet.
Quantitative methods allow them to find patterns that indicate potential writing barriers in the future at a very young age. Mekyska et al.
A random forest model was used to classify children with dysplasia.
The authors include 54 third.
Children in Israel used 10-
Questionnaire on proficiency in Hebrew writing (HPSQ)
Bad writing.
In the adult population, an automated handwritten assessment tool for Parkinson's disease is considered a potential biomarker.
In this work, in order to design a digital diagnostic tool, we build on previous work.
We focused on the clinical relevance of Pediatrics compared to previously determined results.
To this end, we analyzed the data of children with clinical diagnosis to write difficulties and matched it with a group of children with typical development.
We maximize the potential impact of this work by focusing on Latin letters
The most popular script in the world, used by about 2 people.
There are 6 billion people.
In addition, we define features related to features used in current clinical practice.
Our quantitative model utilizes four types of writing features: Geometric aspects of writing, and the use of pressure, tilt, and motion.
We use a random forest classifier to predict the difficulty of writing.
About 96% of the abnormal authors were correctly marked in the Test set (
Really better)
Not less than 1%.
Diagnostic errors in children with dysplasia (
False Yin ratio).
We got an F1. score of 97. 98%.
After establishing the model, we explore and analyze the most important features of diagnostic writing difficulties.
In this analysis, we combine statistical analysis and collaboration with clinicians to exchange examples and comments.
These conclusions are then used to provide insight into the development of a new screening tool that will modernize the current gold
BHK standard test.
Custom message
Chat Online
Chat Online
Chat Online inputting...
Hi, Nice to meet you! I am temporarilly away a momnet, Please can you talk with me by Whatsapp: +86 13582949978 / Skype: eliahe123? Email: info@itatouch.com Thank you in advance! Sincerely, ITATOUCH Sales {{"url":"/about us","text":"Welcome to visit us!"}} Welcome! What can I do for you?
Sign in with: