1Department of Fashion Communication, National Institute of Fashion Technology, India
2Engineer 1, Numerator, India
3Software Development Officer, Digital Health and Care Wales, United Kingdom
BibTex Citation Data :
@article{JMASIF78073, author = {Monika Mehta and Ashmeet Kaur and Madhulika Bhatia}, title = {A Hybrid VGG19-XGBoost Framework with SSIM-Based Feedback for Low-Resource Handwritten Digit Recognition in Digital Learning Systems}, journal = {Jurnal Masyarakat Informatika}, volume = {17}, number = {1}, year = {2026}, keywords = {Handwritten Digit Recognition; Transfer Learning; VGG19; XG Boost; Structural Similarity Index; Educational Technology Feasibility}, abstract = { Handwritten digit recognition is an important component of intelligent educational interfaces, particularly in early learning, digital assessment, and handwriting feedback systems. However, many deep learning-based recognition models rely on large benchmark datasets, whereas practical educational environments may involve limited and heterogeneous learner-generated samples. This study presents a technical feasibility analysis of a hybrid handwritten-digit recognition framework that combines VGG19-based deep feature extraction, XGBoost classification, and Structural Similarity Index Measure (SSIM)-based visual-similarity feedback. A small handwritten-digit dataset was used to simulate a constrained-data setting, with augmentation applied only to the training data to reduce the risk of overfitting. The proposed VGG19-XGBoost pipeline was evaluated against baseline CNN-based models using accuracy, precision, recall, and F1-score, with SSIM used as a supplementary metric to assess the structural similarity between learner input and reference digit forms. The experimental results indicate that the hybrid approach provides stable preliminary classification performance under limited-data conditions and that SSIM can support interpretable visual feedback for handwriting evaluation. However, due to the small number of original samples, the findings should be interpreted as evidence of feasibility rather than as generalizable performance claims. Future work should involve larger real-world datasets, teacher-validated scoring rubrics, and deployment-oriented evaluation in digital learning environments. }, issn = {2777-0648}, pages = {160--175} doi = {10.14710/jmasif.17.1.78073}, url = {https://ejournal.undip.ac.id/index.php/jmasif/article/view/78073} }
Refworks Citation Data :
Handwritten digit recognition is an important component of intelligent educational interfaces, particularly in early learning, digital assessment, and handwriting feedback systems. However, many deep learning-based recognition models rely on large benchmark datasets, whereas practical educational environments may involve limited and heterogeneous learner-generated samples. This study presents a technical feasibility analysis of a hybrid handwritten-digit recognition framework that combines VGG19-based deep feature extraction, XGBoost classification, and Structural Similarity Index Measure (SSIM)-based visual-similarity feedback. A small handwritten-digit dataset was used to simulate a constrained-data setting, with augmentation applied only to the training data to reduce the risk of overfitting. The proposed VGG19-XGBoost pipeline was evaluated against baseline CNN-based models using accuracy, precision, recall, and F1-score, with SSIM used as a supplementary metric to assess the structural similarity between learner input and reference digit forms. The experimental results indicate that the hybrid approach provides stable preliminary classification performance under limited-data conditions and that SSIM can support interpretable visual feedback for handwriting evaluation. However, due to the small number of original samples, the findings should be interpreted as evidence of feasibility rather than as generalizable performance claims. Future work should involve larger real-world datasets, teacher-validated scoring rubrics, and deployment-oriented evaluation in digital learning environments.
Article Metrics:
Last update:
Last update: 2026-05-29 09:36:48
The authors who submit the manuscript must understand that the article's copyright belongs to the author(s) if accepted for publication. However, the author(s) grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. Authors should also understand that their article (and any additional files, including data sets, and analysis/computation data) will become publicly available once published under that license. By submitting the manuscript to Jmasif, the author(s) agree with this policy. No special document approval is required.
The author(s) guarantee that:
The author(s) retain all rights to the published work, such as (but not limited to) the following rights:
Suppose the article was prepared jointly by more than one author. Each author submitting the manuscript warrants that all co-authors have given their permission to agree to copyright and license notices (agreements) on their behalf and notify co-authors of the terms of this policy. Jmasif will not be held responsible for anything arising because of the writer's internal dispute. Jmasif will only communicate with correspondence authors.
Authors should also understand that their articles (and any additional files, including data sets and analysis/computation data) will become publicly available once published. The license of published articles (and additional data) will be governed by a Creative Commons Attribution-ShareAlike 4.0 International License. Jmasif allows users to copy, distribute, display and perform work under license. Users need to attribute the author(s) and Jmasif to distribute works in journals and other publication media. Unless otherwise stated, the author(s) is a public entity as soon as the article is published.
