Blank pages prove to be inefficient in many ways if they are going to be processed, as they do not provide any meaningful information but consume unnecessary storage, processing time, resources and they are difficult to remove using manual intervention and could be error-prone. Hence, an automated approach for blank page detection is essential to improve efficiency in document processing workflows. In this paper MobileNetV2 architecture was used, it’s a lightweight and efficient deep learning architecture used for classifying blank pages. By using transfer learning, the pre-trained MobileNetV2 model is fine-tuned on a custom dataset of blank pages and pages with some handwritten text. This allows the model to detect the characteristics of the blank page and classify the pages accordingly. The proposed approach has several advantages over traditional methods, it is lightweight so it can be run on any device. Experimental results show that this model achieves high accuracy, minimizes false positives, and significantly reduces the manual effort required in document cleaning. Overall, this study highlights the effectiveness of deep learning in addressing the challenge of blank page detection.