Hi,
complaints about using https://en.wikipedia.org/wiki/Lenna appeared recently against multiple pieces of software and multiple distributions. Could you please replace demo/lena.rgb with a freely redistributable image? It is used in rgbalpha.dem and circles.dem. Both demos are meaningful but could work with almost any image equally well. I can pick some image and provide a patch if you wish.
If you can legally distribute Lenna, please provide some background. (I vaguely remember an event a few years ago where the company unofficially promised not to accuse anybody for using the picture. But it probably does not make the picture freely redistributable. I Am Not A Lawyer.)
Thank you very much,
Fero
If you want to pursue this, please prepare a patch and/or replacement demos.
So far as I am aware, recent complaints about use of the Lena Söderberg image are not rooted in copyright/licensing issues but rather on distaste for its provenance (Playboy Magazine).
Replaced lena with tux. The image is already used in other demos and I really hope it is safe. The circles placed just with mouse clicking, not sure if they were generated somehow automatically.
No, sorry. I don't think the Tux image is a suitable replacement at all. It preserves essentially none of the properties that make the Lena image a nice graphics test. Computer-generated images in general are quite different from photographic images. The only other standard test image roughly similar in frequency of use to the Lena image is the Mandrill image. The "red and green peppers" image is maybe a distant third. These three are from the USC/SIPI collection of test images. But the Mandrill is not a good example of smooth shading, and the colors are IMHO too garish. The color palette in the peppers image is biased to red/green.
A ideal replacement image should be
1) photographic
2) illustrate smooth shading and a reasonably broad color palette (Lena is not ideal for the latter)
3) contain sufficient edge-content to allow computation of edges and nodes as in the existing demo "circles placed with mouse clicks" are not at an adequate replacement for a demo that illustrates quantitative detection or display of features in the image.
4) shared as a test image with other common graphics programs, so that comparison is easier
5) decent quality at 512x512 resolution so that it doesn't take up too much space in the demo collection
I really have not come across many standard test images (point 4), but even if we were to relax this point it would be good to keep the other points in mind.
I obviously underestimated and misunderstood the purpose of this demo. I did not know that gnuplot is able to do quantitative detection of edges. How were original points calculated?