|
LSTM-Based Autoencoders in Online Handwriting Data Augmentation and Preprocessing
Gavenciak, Michal
On-line handwriting analysis is a research field that is among others used in assessment of handwriting difficulties (HD), which can be manifestations of degenerative brain diseases such as Parkinson’s disease in the elderly, or developmental dysgraphia in children. Using advanced modelling approaches or artificial intelligence is often difficult because of the limited data availability in both demographic cohorts. In this article, a data processing approach, using LSTM-based autoencoders, is described as a way of augmenting the database with semisynthetic data or preprocessing the data to improve the performance of feature-based classification. The proposed method has led to a 3 percentage point increase in classification accuracy when compared to baseline. While the improvement is marginal, it highlights another possible area of research to improve the efficacy of automated HD assessment.
|
|
Research of Advanced Online Handwriting Analysis Methods with a Special Focus on Assessment of Graphomotor Disabilities in School-aged Children
Zvončák, Vojtěch ; Havigerová,, Jana Marie (oponent) ; Drotár,, Peter (oponent) ; Mekyska, Jiří (vedoucí práce)
Graphomotor abilities (GA) represent a set of psychomotor activities that are executed during drawing and writing. The GA are prerequisites for mastering of elementary school skills, particularly writing. Children in 1 st and 2 nd grade can experience difficulties in execution of simple graphomotor tasks (GD) and later in 3 rd and 4 th grade even in handwriting tasks (HD). The disruption of processes involved in handwriting is generally called Developmental Dysgraphia (DD). The prevalence of DD in the Czech Republic is 3–5 %. To this day the DD is evaluated subjectively by teams of psychologists and special educationalist. Currently, an objective measuring tool that asses the properties of GD or HD is missing in practice. Consequently, this thesis is aiming to identify symptoms associated with graphomotor disabilities in school–aged children and design new parameters quantifying them. For this purpose, a new complex GA protocol was proposed (36 tasks), which represents an environment, where the identified symptoms can be manifested (24 symptoms). Moreover, 76 quantifying features were introduced. A new graphomotor difficulties rating scale (GDRS) was designed based on computerised analysis of handwriting. Finally, new online handwriting parameters based on advanced signal processing techniques were designed and tested, which can assess poor dexterity or unspecified motor clumsiness. The GDRS represents a novel and modern objective measurement tool, that is not yet available in the Czech Republic or in other countries. Its utilization will help in the modernization of DD diagnosis and in the remediation process. With proper research, it could be adapted into other languages as well. Moreover, the methodology can be used and optimized for other diseases, which affects GA, such as Autism, Attention Deficit Coordination Disorder (ADHD) or Developmental Coordination Disorder (DCD).
|
|
Exploring the Contribution of Isochrony-based Features to Computerized Assessment of Handwriting Disabilities
Gavenčiak, M. ; Zvončák, V. ; Mekyska, J. ; Šafárová, Katarína ; Čunek, Lukáš ; Urbánek, Tomáš ; Havigerová, Jana Marie ; Bednářová, Jiřina ; Galáž, Z. ; Mucha, J.
Approximately 30–60 % of the time children spend in school is associated with handwriting. However, up to 30 % of them experience handwriting disabilities (HD), which lead to a decrease in their academic performance. Current HD assessment methods are not unified and show signs of subjectivity which can lead to misdiagnosis. The aim of this paper is to propose a new approach to objective HD assessment based on the principle of movement isochrony. For this purpose, we used a database of 137 children attending a primary school, who performed a transcription and dictation task, and who were associated with a BHK (Concise Evaluation Scale for Children's Handwriting) score. Employing a machine learning model, we were able to estimate this score with 18 % error. An interpretation of the model suggests that the isochrony-based features could bring new benefits to the objective assessment of HD.
|
|
Research of Advanced Online Handwriting Analysis Methods with a Special Focus on Assessment of Graphomotor Disabilities in School-aged Children
Zvončák, Vojtěch ; Havigerová,, Jana Marie (oponent) ; Drotár,, Peter (oponent) ; Mekyska, Jiří (vedoucí práce)
Graphomotor abilities (GA) represent a set of psychomotor activities that are executed during drawing and writing. The GA are prerequisites for mastering of elementary school skills, particularly writing. Children in 1 st and 2 nd grade can experience difficulties in execution of simple graphomotor tasks (GD) and later in 3 rd and 4 th grade even in handwriting tasks (HD). The disruption of processes involved in handwriting is generally called Developmental Dysgraphia (DD). The prevalence of DD in the Czech Republic is 3–5 %. To this day the DD is evaluated subjectively by teams of psychologists and special educationalist. Currently, an objective measuring tool that asses the properties of GD or HD is missing in practice. Consequently, this thesis is aiming to identify symptoms associated with graphomotor disabilities in school–aged children and design new parameters quantifying them. For this purpose, a new complex GA protocol was proposed (36 tasks), which represents an environment, where the identified symptoms can be manifested (24 symptoms). Moreover, 76 quantifying features were introduced. A new graphomotor difficulties rating scale (GDRS) was designed based on computerised analysis of handwriting. Finally, new online handwriting parameters based on advanced signal processing techniques were designed and tested, which can assess poor dexterity or unspecified motor clumsiness. The GDRS represents a novel and modern objective measurement tool, that is not yet available in the Czech Republic or in other countries. Its utilization will help in the modernization of DD diagnosis and in the remediation process. With proper research, it could be adapted into other languages as well. Moreover, the methodology can be used and optimized for other diseases, which affects GA, such as Autism, Attention Deficit Coordination Disorder (ADHD) or Developmental Coordination Disorder (DCD).
|
|
Semi-Automatic Segmentation Of On-Line Handwriting
Gavenčiak, Michal
This paper deals with the automation of digital trace data segmentation. The data are obtained from a digitizing tablet and are then subjected to handwriting analysis, providing quantified information about a person’s handwriting, which might help in the diagnosis of handwriting difficulties. In order to successfully analyze the data, they must be segmented by individual handwriting exercise. Implementation of a python-based program with a GUI is described along with its basic functionality.
|