Convolutional Neural Networks (CNNs) have been introduced as a reliable deep learning solution for tasks related to pattern recognition, especially in the area of visual data processing. They have been demonstrated to be robust in accurately recognizing scribes from writing documents. In this paper, we present a CNN-based technique to extract and characterize the unique writing features of individuals. Our approach processes handwritten documents by segmenting them into smaller images, each of which represents a coherent component. These are then processed by our proposed CNN model, IDWriter, to identify the writers. IDWriter operates as a fully integrated network, specifically trained on these connected writing segments. To enhance identification accuracy, a decision-classification mechanism is introduced, which averages the model’s predictions across the different components. Our proposal achieves promising results compared to the literature, as evidenced by experimental results on five benchmark datasets.