Evaluation and visual representation of three-dimensional images
With growing interest I read a paper about an actually dry subject, the XISF Version 1.0 Specification (DRAFT 9.3),
http://pixinsight.com/doc/docs/XISF-1.0-spec/XISF-1.0-spec.html Especially the topic of three-dimensional images aroused my interest:
In section 5, "Overview of XISF 1.0 Features", the dimensionality and the number of channels of an image are defined as distinct conceptions. In section 8.5.1 "Structure and Properties" the concept is explained in more detail, and Figure 1 - "Structure of a two-dimensional image" illustrates the case of a two-dimensional image. The second paragraph of section 8.5.2 clearly defines three-dimensional images and section 11.5.1 among other things deals with the image attribute 'geometry'.
According to this specification Pixinsight is (or will be?) ready to represent three-dimensional images internally, but as far as I am aware of, Pixinsight by now does not make use of it. I would like to have advanced processes in Pixinsight enabling to open, save, evaluate and visually represent the data of three-dimensional images. Such functionality is commonly applied in image processing in the field of e. g. medicine or biology. However, also in astrophotography one can view a stack of dark frames or a stack of light frames as a three dimensional image and I think it's worthwhile giving thought to it:
As an example, in the examination of a camera sensor I needed functionality that would let me examine a specific hot pixel (x, y) and output (as chart and as numerical data) the intensity as function of the depth (z coordinate) in a stack of dark frames.
Second example: imagine a stack of images as a cuboid. Normally, as in the case of the blink process, we look at it in a way that we see the (x, y) image of a specific z coordinate. The x axis is the direction of rows and the y axis is the direction of columns of the image. Even if blink treats the inspected images as separate, they altogether could be considered as one three-dimensional image. When we scroll through them, we virtually change the z coordinate. Now I would like to be able to (graphically) turn the cuboid in space and view it from a different direction. So you could visually represent two-dimensional (x, z) or (y, z) images and scroll through the y coordinate or the x coordinate, respectively. In this way one could e. g. rapidly and unambiguously identify cosmic ray hits.
Juan Conejero wrote in topic 4529 (August 2012): "PixInsight is an image processing and analysis platform specialized in astronomy and other technical imaging fields." Advanced functionality regarding three-dimensional images would facilitate specific evaluations and visual representations in astronomy and make Pixinsight more interesting for a broader public (potential users in other scientific fields) too.
I guess that internally in Pixinsight the needed processes are already at hand. Unfortunately my programming skills are not sufficient, otherwise I now had an interesting programming project for a JavaScript. In any case, I am curious about the further development of Pixinsight and specifically about the support of three-dimensional images. Perhaps the developers can take the time for commenting on their point of view?
Bernd
