pixel resolution notes v0.4

For all these examples, assume the axes are :

[
    {"name": "y", "type": "space", "unit": "micrometer"},
    {"name": "x", "type": "space", "unit": "micrometer"}
]

We will represent an image with

• a pixel resolution of 0.1 in y and 0.3 in x,
• two scale levels, where scale level "1" is downsampled by a factor of two relative to "0".

Example 1

How I would do it - every scale level's transform describes its physical resolution.

{
    "datasets": [
        {
            "path": "0",
            "coordinateTransformations": [{ "type": "scale", "scale": [0.1, 0.3] }]
        },
        {
            "path": "1",
            "coordinateTransformations": [{ "type": "scale", "scale": [0.2, 0.6] }]
        }
    ]
}

Example 2

Another valid and reasonable way to do it. "datasets" transformations are relative to the highest scale, and the "outer" transformation encodes resolution.

{
    "datasets": [
        {
            "path": "0",
            "coordinateTransformations": [] // equivalent to the identity
        },
        {
            "path": "1",
            "coordinateTransformations": [{ "type": "scale", "scale": [2, 2] }]
        }
    ],
    "coordinateTransformations": [{ "type": "scale", "scale": [0.1, 0.3] }],
}

Example 3

This is equivalent to the above, but please don't do this!!!

{
    "datasets": [
        {
            "path": "0",
            "coordinateTransformations": [{ "type": "scale", "scale": [-0.01, 3.0] }]
        },
        {
            "path": "1",
            "coordinateTransformations": [{ "type": "scale", "scale": [-0.02, 6.0] }]
        }
    ],
    "coordinateTransformations": [{ "type": "scale", "scale": [-10, 0.1] }],
}