i have 3 bugs with Spot detector and ICY

[AMAR Majid]

hello, I use frequently ICY and Spot detector,

and I see 3 bugs

1) after select the ROIs, ans after starting spot detector, the software found spots out side of my selection.

   (see the attached file 1. the superposition Icy Bug)

2) to eliminate this problem, I start doing a size filtring (0 to 25 pixels size) to eliminate the artecafcts,

but the softaware keeps some spots biger than 25 pixels

  (see the attached file 2. bug size filtre)

3) the excel output file after the spot detector,

for exemple: I do in the ordre 4 ROIs (Cortex frontal 1, after Cortex frontal 2, after Cortex frontal 3 and to finish the Union of the cortex frontal)

and in the output excel file we found everything in the disorder: cortex frontal 1, union cortex frontal, cortex frontal 3 and cortex frontal 2

  (see the attached file 3. bug classement ROI)

execpt those 3 bugs, ICY and spot detector are a powerful tool

[AMAR Majid]

the attached files

[Fabrice de Chaumont]

Hi Majid,

The detection is considered inside an ROI considering its center of mass. In your case, the detection is completely outside the ROI, but as it is a big convex object, the center of mass is in the ROI. This is why it is kept. Well, this is my bad anyway. You can filter by size to remove those very big objects as you suggested in your point #2. I am about to publish a new version of the spot detector where the filtering (and many other things) are fixed.

The bug #3 is also a kernel problem. When the spot detector query Icy to get the ROI list, the ROI can be in any order.
I am discussing with Stephane to see if sorting them with the current selected criteria in the ROI tab would be great, and by creation time by default. I will check this.

I understand it's a problem to have it random.

>xecpt those 3 bugs, ICY and spot detector are a powerful tool

if you like/dislike it, rate it !! ==> http://icy.bioimageanalysis.org/plugin/Spot_Detector

Best,

Fabrice

[Fabrice de Chaumont]

Hi Majid,

I just released a new version of the spot detector.

best,

[AMAR Majid]

Hi Fabrice, 

thank you for your answer,

I just did a small test of Spot detector,

I see a lot of modification in the spot detector,

normaly the bug number 1, and number 2 were solved,

concerning the third one - the classement of ROI in the Excel out put file - the order of the ROIs does not correspond to how I ordered them in my selection.

I just did: ROI 1, after 2, after 3, and finelly the Union of 1+2+3

and in the excel: I found: 1, Union, 3 and 2

realy thank you, for this free and powerfull tool.

and a special thanks for your fast response.

Majid

[Fabrice de Chaumont]

Hi Majid,

Thank you for your message.

The ROI ordering is more a problem of the kernel. We are working on it with Stephane.
What you can do is naming the ROIs, ( by clicking on the name of the ROI) so you will recover (and maybe sort) them in the excel file.

Anyway thank you for rising this problem, we will work on it.

Best,

Fabrice