It sometimes arise that I need to use, for an article, a book chapter or any other written work, figures that I have already published (not figures published by others). These are usually schematic figures explaining what a chemical or physical system look like, how an algorithm works, etc. Thus, instead of reusing the exact same figure, I can easily create another one, conveying the same message with (more or less subtle) differences in presentation.
However, I have no idea how to answer the following question: how different need two figures be to avoid the second one infringing on the copyright of the first? What is a good rule of thumb to be used? Is it enough to change one of the following:
- color scheme
- viewpoint of a 3D visualization
- moving around blocks in a 2D diagram (or mind map)
- changing axis properties (labels, tick mark positions, etc.) in a graph
In general, I would argue that the new plot needs to do one of two things:
- present materially different content relative to the old image
- present the same material in a different context.
Graphs are generally visual representations of numerical data. The numerical data itself cannot be copyrighted (because the data are "facts"), only the presentations thereof. However, simply changing axis labels or colors, and other similar "gloss" changes, don't really make the graph any different. Rotating a 3D graph, or changing significantly the relationship of blocks in a mind map or diagram to show off different features would represent a change in the message and presentation, and therefore would be OK.
However, chemical structures and mathematical equations are not normally considered subject to copyright, as they represent "facts," and cannot be arbitrarily drawn or represented; there's only so many "legal" ways to write them. That said, cutting and pasting somebody's figure from another article would be a copyright violation; recreating it yourself would not. (Otherwise, we would never be able to write E = mc^2 without it being a copyright violation!)
If there are doubts, however, you can always try to contact the journal in question. If needed, you can ask for permission to reuse a figure, especially if it's one you've already created. Most journals already permit self reuse, so long as you include an appropriate credit and citation of the original.