search

DrHanLim / EzColocalization

GNU General Public License v3.0
10 stars 2 forks source link

FIX: resolve Show Threshold and Roi name issue in the new ImageJ verions #10

Closed shenghuanjie closed 3 years ago

shenghuanjie commented 3 years ago

Fixes #8 Fixes #9

Change file from CRLF to LF Remove unused imports and reorganize imports

shenghuanjie commented 3 years ago

@westonstauffer. Please give a thumb up if it is working for you.

westonstauffer commented 3 years ago

ROIs in manager now have cell numbers but still have duplicates. For example: Image 1 cell 1 Image1cell1 Image 1 cell 2 is a duplicate of Image 1 cell 1 Image1cell2 Image 1 cell 3 is the true cell 2 in this example Image1cell3

westonstauffer commented 3 years ago

Cell filters other than the pre-watershed filter identify additional incorrect cells for analysis as in this example:

Without cell filters. Yellow outline shows selected cells No_filters

With pre-watershed filter of 500-Infinity, two cells with >500 pixels are selected and shown in analysis table pre-watershed_filter

With cell filter 1 set to "Area" and 500-Infinity there are additional cells shown with yellow outline, and additional cells in data table cell_filter_1

shenghuanjie commented 3 years ago

ROIs in manager now have cell numbers but still have duplicates. For example:

The following change is just a disaster waiting to happen. ImageJ should not have decreased the visibility of part of a protected function. What happened recently is the update to RoiManager got separated into a private function from the protected saveResults we override in EzColocalization. After this change, there is no way for us to have control over the update happening in RoiManager since the new addToRoiManager is a private method. https://github.com/imagej/imagej1/blob/ba4a8ee10f3cbe06393f4046a88bfa67683aedfd/ij/plugin/filter/ParticleAnalyzer.java#L910-L912

shenghuanjie commented 3 years ago

Heat maps scaled by "image" and "stack" do not clear outside of ROIs and are identical.

Are you sure we were clearing the outside of ROIs before for "image" and "stack"?

westonstauffer commented 3 years ago

Heat maps scaled by "image" and "stack" do not clear outside of ROIs and are identical.

Are you sure we were clearing the outside of ROIs before for "image" and "stack"?

You are right. We didn't clear outside of ROIs in "image" or "stack" scaling. I have found that the "stack" scaling is scaling by the first image in the stack. In this example, Image 1 in a stack has 1/5 its normal intensity and for "cell" and "stack" scaling looks similar: "image" scaling Imagescale_1:5intensity "stack" scaling Stackscale_1:5intensity

But other images are shifted to mostly maximum values for the "stack" scaling because image 1 had such a lower range for scaling "stack" scaling Stackscale "image" scaling Imagescale