Prediction of Grade, Gender, and Academic Performance of Children and Teenagers from Handwriting Using the Sigma-Lognormal Model
Authors: Adrian Iste, Kazuki Nishizawa, Chisa Tanaka, Andrew Vargo, Anna Scius-Bertrand, Andreas Fischer, Koichi Kise
Organizations: Osaka Metropolitan University, Osaka, Japan · University of Applied Sciences and Arts Western Switzerland, Fribourg, Switzerland · University of Fribourg, Fribourg, Switzerland · German Research Center for Artificial Intelligence, Japan Laboratory, Osaka, Japan
Digital handwriting acquisition enables the capture of detailed temporal and kinematic signals reflecting the motor processes underlying writing behavior. While handwriting analysis has been extensively explored in clinical or adult populations, its potential for studying developmental and educational characteristics in children remains less investigated. In this work, we examine whether handwriting dynamics encode information related to student characteristics using a large-scale online dataset collected from Japanese students from elementary school to junior high school. We systematically compare three families of handwriting-derived features: basic statistical descriptors of kinematic signals, entropy-based measures of variability, and parameters obtained from the sigma-lognormal model. Although the dataset contains dense stroke-level recordings, features are aggregated at the student level to enable a controlled comparison between representations. These features are evaluated across three prediction tasks: grade prediction, gender classification, and academic performance classification, using Linear or Logistic Regression and Random Forest models under consistent experimental settings. The results show that handwriting dynamics contain measurable signals related to developmental stage and individual differences, especially for the grade prediction task. These findings highlight the potential of kinematic handwriting analysis and confirm that through their development, children's handwriting evolves toward a lognormal motor organization.
Figures & tables
Figure 1: Distribution of the ages of the students in Wacom dataset.
Grade
Boys
Girls
Total
1
27
23
50
2
24
26
50
3
23
27
50
4
17
30
47
5
22
28
50
6
30
20
50
Table 1: Distribution of students across grades.
Figure 2: Examples of drills from subject “Japanese language” for grade 1 to 9.
Figure 3: Distribution of number of questions in the drills of the subject Japanese language.
Figure 4: Score distributions analysis in Wacome dataset
Feature
Description
Number of Lognormal Components ( C )
The number of lognormal components required to model each stroke. This parameter reflects the degree of motor segmentation.
Velocity Profile Signal-to-Noise Ratio (SNR)
Signal-to-noise ratio of the velocity profile, which quantifies how well the reconstructed lognormal model explains the observed kinematic signal.
Normalized SNR (SNR/ C )
Signal-to-noise ratio divided by the number of lognormal components, normalizing reconstruction quality by model complexity.
Time Dispersion ( D )
The lognormal time dispersion parameter controlling the temporal spread of each motor impulse. It is associated with the duration and fluidity of the movement.
Time Delay ( t0 )
The onset timing of motor commands within the stroke sequence.
Location Parameter ( μ )
Determines the temporal positioning of each motor impulse within the stroke.
Table 2: Description of Lognormal Stroke Features
Figure 5: Prediction of the grade for each set of feature using Linear Regression (left) and Random Forest (right) using 5-fold cross validation. One point corresponds to a prediction of a point from the test fold.
Figure 6: Box plot of the signal-to-noise ratio over the number of lognormal components across all grades.
Figure 7: Confusion matrix of the gender classification for each set of feature using Logistic Regression (top) and Random Forest (bottom) with 5-fold cross validation.
Feature Set
Logistic Regression
Random Forest
Majority-class baseline
ACC
F1
AUC
ACC
F1
AUC
Majority ACC
Base
0.611
0.535
0.626
0.600
0.551
0.625
0.537
Entropy
0.552
0.445
0.574
0.552
0.493
0.595
0.537
Sigma-lognormal
0.653
0.604
0.725
0.658
0.614
0.693
0.537
Table 3: Gender classification performance over all grades using 5-fold cross validation for all sets of feature compared to a majority-class baseline.
Figure 8: Confusion matrix of the prediction of the academic performance of a student for each set of feature using Logistic Regression (top) and Random Forest (bottom) with 5-fold cross validation.
Feature Set
Logistic Regression
Random Forest
Baseline
ACC
F1
AUC
ACC
F1
AUC
Majority ACC
Base
0.596
0.569
0.645
0.579
0.540
0.624
0.526
Entropy
0.607
0.583
0.642
0.592
0.560
0.641
0.526
Sigma-lognormal
0.628
0.607
0.673
0.615
0.579
0.665
0.526
Table 4: Classification results over all grades of the overall academic performance of a student using 5-fold cross validation for all sets of feature compared to a “always majority” model.
While handwriting has traditionally been studied for character recognition and disease classification, its potential to reflect day-to-day physiological fluctuations in healthy individuals remains unexplored. This study examines whether daily variations in sleep-related recovery states can be inferred from online handwriting dynamics. % We propose a personalized binary classification framework that detects low-recovery days using features derived from the Sigma-Lognormal model, which captures the neuromotor generation process of pen strokes. In a 28-day in-the-wild study involving 13 university students, handwriting was recorded three times daily, and nocturnal cardiac indicators were measured using a wearable ring. For each participant, the lowest (or highest) quartile of four sleep-related metrics -- HRV, lowest heart rate, average heart rate, and total sleep duration -- defined the positive class. Leave-One-Day-Out cross-validation showed that PR-AUC significantly exceeded the baseline (0.25) for all four variables after FDR correction, with the strongest performance observed for cardiac-related variables. Importantly, classification performance did not differ significantly across task types or recording timings, indicating that recovery-related signals are embedded in general movement dynamics. These results demonstrate that subtle within-person autonomic recovery fluctuations can be detected from everyday handwriting, opening a new direction for non-invasive, device-independent health monitoring.
Chisa Tanaka, Andrew Vargo, Anna Scius-Bertrand +2
Osaka Metropolitan University, Osaka, Japan · University of Applied Sciences and Arts Western Switzerland, Fribourg, Switzerland · University of Fribourg, Fribourg, Switzerland +1
Dysgraphia is a specific learning disability that is prevalent among school-age children. It affects handwriting coherence, quality, fluency, and legibility, often hindering academic achievement and early learning development. This motor coordination disorder is typically diagnosed through subjective assessments based on clinician observation, which can be timeconsuming and prone to variability. In this paper, we introduce a deep learning-based framework for objective dysgraphia detection using online handwriting data captured via digitizing tablets. The proposed framework relies on two complementary branches: the first pipeline extracts both handcrafted and embedding-based kinematic features directly from raw temporal signals, while the second leverages image-based representations of the temporal signals generated using continuous wavelet transforms (CWT) and Gramian Angular Fields (GAF). The resulting features are then fused to leverage the complementary strengths of both representations. The four representations were evaluated separately and jointly using the publicly available DiaGraMo dataset, showing that the fusion of GAF, MOMENT, and hand-crafted kinematic features outperforms each individual representation, as well as other fusion schemes. These findings highlight the potential of the complementarity of image and signal based representations for more objective dysgraphia detection.
Lydia Ouhib, Yassine Ouzar, Zoé Pinseel +2
LIASD · LIASD Laboratory, University of Paris 8, Saint-Denis, France · Centre Jacques Calv´e, Fondation Hopale, Berck, France
Digital pens are commonly used to write on digital devices, providing the handwriting trace and enhancing human-computer interation. This study focuses on a digital pen equipped with kinematic sensors, allowing users to write on any surface while simultaneously preserving a digital trajectory of handwriting. This technology holds significant potential as a valuable educational tool, particularly in classrooms where it can facilitate the process of learning to write. A major issue is based on the difference in captured signals between adults and children. For similar handwriting trace, we have large differences in sensor signals due to differences in speed and confidence in the handwriting gesture of children. To address this, we investigate a domain adaptation approach to build a unified intermediate feature representation aimed at facilitating the trajectory reconstruction. We demonstrate the interest of domain adaptation methods in leveraging existing knowledge for application in different contexts. Specifically, we compare our domain adaptation approach with two other methods: training the model from scratch and fine-tuning the model.
Florent Imbert, Romain Tavenard, Yann Soullard +1
IRISA, Universite de Rennes, INSA Rennes, Rennes, France · IRISA, Universite Rennes 2, Rennes, France