Abstract: We present a new dataset for form understanding in noisy scanned documents (FUNSD) that aims at extracting and structuring the textual content of forms. The dataset comprises 199 real, fully ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results