A Gated and Bifurcated Stacked U-Net Module for Document Image Dewarping
This addresses the problem of removing distortions from document images captured with handheld devices, which is an incremental improvement in a specific computer vision domain.
The paper tackles document image dewarping by proposing a Gated and Bifurcated Stacked U-Net module that predicts a dewarping grid to create distortion-free images, achieving state-of-the-art performance on the DocUNet dataset while using only 8% of the training data compared to previous methods.
Capturing images of documents is one of the easiest and most used methods of recording them. These images however, being captured with the help of handheld devices, often lead to undesirable distortions that are hard to remove. We propose a supervised Gated and Bifurcated Stacked U-Net module to predict a dewarping grid and create a distortion free image from the input. While the network is trained on synthetically warped document images, results are calculated on the basis of real world images. The novelty in our methods exists not only in a bifurcation of the U-Net to help eliminate the intermingling of the grid coordinates, but also in the use of a gated network which adds boundary and other minute line level details to the model. The end-to-end pipeline proposed by us achieves state-of-the-art performance on the DocUNet dataset after being trained on just 8 percent of the data used in previous methods.