PDFFigure2 might do better job in separating figures compared to the current model

[X-Bruce-Y]

Hi VikParuchuri, I'm really grateful that you share such a nice tool. I happen to find in one case that PDFFigure2 does better in seperating figures, using the same pdf https://www.thelancet.com/journals/ebiom/article/PIIS2352-3964(22)00586-2/fulltext (open access, downloadable). Specifially, figure 5 got chunked differentially.

---

marker (into 2 pieces, and overall missing one part from the original figure)

and

PDFFigure2

---

I'm no expert and I used PDFFigure2 through https://github.com/MuiseDestiny/zotero-figure.

It would be nice if you find out something useful there.

Cheers,
Bruce

[X-Bruce-Y]

Forgot to mention, marker also extracts tiny icons in this example, which is not favorable. It would be better if there could be a flag to screen images (based on size, etc).