Scholarlink

The current study is classification of dysgraphia from handwritten images using an ensembled Super-Resolution Generative Adversarial Networks (SRGANs), Convolutional Neural Networks (CNNs), and Long Short-Term Memory networks (LSTMs). The proposed model aims to address the challenges of dysgraphia classification by leveraging the high-resolution image generation capabilities of SRGAN, the feature extraction capabilities of CNNs, and the sequential modeling capabilities of LSTMs. The AI model is trained using primary set of handwriting samples collected from individuals with and without dysgraphia 101 images. We first employ SRGAN to enhance the resolution of input images, thereby improving the quality of features extracted by subsequent layers. Next, a CNN architecture is utilized to extract discriminative features from the enhanced images. Finally, an LSTM network is employed to capture the temporal dependencies in the sequence of extracted features, enabling the model to effectively classify dysgraphia. The performance model was evaluated using standard metrics such as precision, recall, and F1-score. The results of the study have revealed that the ensembled model achieved a high classification accuracy for dysgraphia, indicating its potential usefulness as a diagnostic tool for dysgraphia.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Hybrid Model of SRGAN CNN-LSTM for Classifying Dysgraphia in Primary School Children Using Handwritten Images

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Sailaja Mulakaluri,
  • G. S. Girisha

摘要

The current study is classification of dysgraphia from handwritten images using an ensembled Super-Resolution Generative Adversarial Networks (SRGANs), Convolutional Neural Networks (CNNs), and Long Short-Term Memory networks (LSTMs). The proposed model aims to address the challenges of dysgraphia classification by leveraging the high-resolution image generation capabilities of SRGAN, the feature extraction capabilities of CNNs, and the sequential modeling capabilities of LSTMs. The AI model is trained using primary set of handwriting samples collected from individuals with and without dysgraphia 101 images. We first employ SRGAN to enhance the resolution of input images, thereby improving the quality of features extracted by subsequent layers. Next, a CNN architecture is utilized to extract discriminative features from the enhanced images. Finally, an LSTM network is employed to capture the temporal dependencies in the sequence of extracted features, enabling the model to effectively classify dysgraphia. The performance model was evaluated using standard metrics such as precision, recall, and F1-score. The results of the study have revealed that the ensembled model achieved a high classification accuracy for dysgraphia, indicating its potential usefulness as a diagnostic tool for dysgraphia.