Automated Table Detection Using R-CNN: A Cloud-Based Approach

In today's data-driven world, there is an ever-increasing requirement for effective data processing from document images that contain tables. These documents contain a wide variety of complexity, such as different layouts and visual distortions. This paper presents solution for table detection from document images by proposing deep learning models such as RCNN and MobileNetV2. The deep learning models were enhanced by intensive hyper-parameter tuning to optimize model parameters. In this research work experimentation is done on the Table bank dataset to demonstrate model performance. In addition to this, the paper emphasizes on practical deployment through Microsoft Azure. The work presented here represents a substantial improvement in the process of automating the detection of tables inside document pictures.