A theoretical justification of warping generation for dewarping using CNN. (January 2021)

• Record Type: Journal Article
• Authors: Garai, Arpan; Biswas, Samit; Mandal, Sekhar
• Is Part Of: Pattern recognition. Volume 109(2021)
• Publication Date: 2021-01
• Subjects: Dewarping -- Artificial neural netwroks -- Synthetic image generation; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2020.107621
• Languages: English
• ISSNs: 0031-3203

Abstract

Highlights: A mathematical model for warping generation is proposed. The model uses geometric parameters like the camera angle, camera position, distance from the object to camera and curvature the surface to generate warping. Synthetic images are generated using the model. Value of the geometric parameters is estimated using CNN from a 2D warped image. Performances of both the synthetic image generation and dewarping model are analysed. Abstract: Dewarping is a necessary preprocessing step to recognize text from a distorted camera captured document image. According to recent literature, deep learning-based approaches perform with higher accuracy in similar domains. The deep learning-based neural networks are not yet fully explored in the domain of dewarping. To fill this gap, we propose a dewarping approach based on the convolutional neural network. A large number of images are required to train such networks. However, it is a tedious job to capture such a large number of images. Hence, it is required to generate synthetic warped images for the training phase of the deep learning-based neural network. The existing synthetic warped image generation methods are heuristic-based. In this paper, we propose a novel mathematical model for the generation of warped images. The proposed model takes some parameters such as depth of the surface, camera angle, and camera position and generates the corresponding warped image. These parameters are the ground truth for that particular warped image. We use a Convolutional Neural Network (CNN) based model to estimate the warping parameters from a 2 D warped image for dewarping. In the training phase of CNN based model, the synthetic images and their corresponding ground truth are used. Next, the trained model is used to dewarp the unknown warped images. The performance of the proposed warping model is analyzed. Finally, the proposed dewarping method is compared with existing approaches. In both cases, the results are encouraging.