This is related to #698
Some documents have main figures and supplementary figures.
If in those cases, the segmentation model labels the supplementary figures as annex,
then the content is passed separately to the fulltext model.
If the fulltext then correctly labels it as figure, then the figures from the annex are not included in the output.
This is related to #698
Some documents have main figures and supplementary figures.
If in those cases, the segmentation model labels the supplementary figures as
annex,then the content is passed separately to the
fulltextmodel.If the
fulltextthen correctly labels it asfigure, then the figures from theannexare not included in the output.