Hello I am reporting a bug I found in the Bioformat Importer function. When the selected file is opened (regardless of type TIFF, JPEG, ect.) and a resolution that is selected other than the original, the scale of each physical size is not adjusted. For example, as shown below, the original resolution has each pixel in a 15504 by 14263 as .252 microns. However, when converted to a 834x768 resolution the physical size does not adjust accordingly, it remains .252 microns. I fixed this by setting the scale of each pixel to adjust for the change in resolution, in this case I had each pixel equal to (.252*(14263/768))=4.68. This adjustment then resulted in outputs at the given resolution to accurately match the test images I put in.
Welcome to the forum, @sbaygani!
How are you resizing the image? Using Image Adjust Size…?
I tested by opening the Bridge sample, setting unit of length to
feet, setting the value to pixel width/height/depth to
0.1, then running Image Adjust Size…. Afterward, the physical pixel size (as shown in the Image Properties…) dialog adjusted accordingly, to
0.2 width and height, such that the image status still said 51.2 feet x 51.2 feet.
This Bug is a problem for me too.
It occurs for the following image: (The file contains a pyramid / multiple resolutions of the same image)
It’s possible to open different downscaled versions of the image with ImageJ & Bioformats.
But ImageJ pretends the physical pixel size is always the same, independent of th scaling of the image.
The bug is propably inside Bioformats.
The wrong physical pixel sizes can be found in the OME Metadata, provided directly by the Bioformats API.
This does indeed appear to be a bug within Bio-Formats. We have an existing Trello card (https://trello.com/c/INs2Dw2I/147-incorrect-physical-pixel-sizes-with-pyramids) which raised a similar issue. I will update that existing card as the reports listed here suggest that the issue is independent of format type.