Abstract: We propose a high-performance fully convolutional neural network (FCN) for historical document segmentation that is designed to process a single page in one step. The advantage of this model ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback