Several members of the team have reported it would be useful for their own readings of papers to be able to jump from a figure to the place where that figure is discussed in the text, which can be "a real pain".
To support this feature, we would need to use some coloring technique like before (maybe for figures, setting background color instead of text color) to detect the positions of figures, and also the references to those (that could probably be done with the typical coloring technique). The design of the extraction method would be quite similar to how we match citations and their reference information.
We could also extract metadata about figures (e.g., their captions) that we could show in a tooltip before someone jumps to the figure (though I imagine we may just want to show the figure itself, including the caption).
andrewhead
commented
4 years ago
Several members of the team have reported it would be useful for their own readings of papers to be able to jump from a figure to the place where that figure is discussed in the text, which can be "a real pain".
To support this feature, we would need to use some coloring technique like before (maybe for figures, setting background color instead of text color) to detect the positions of figures, and also the references to those (that could probably be done with the typical coloring technique). The design of the extraction method would be quite similar to how we match citations and their reference information.
We could also extract metadata about figures (e.g., their captions) that we could show in a tooltip before someone jumps to the figure (though I imagine we may just want to show the figure itself, including the caption).