ADCI Wiki

User Tools

Site Tools

Trace:
main:presetimgselectionmodels

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Next revision Both sides next revision
main:presetimgselectionmodels [2017/03/31 14:57]
127.0.0.1 external edit 		main:presetimgselectionmodels [2017/05/03 14:22]
bshirley
Line 2: Line 2:
 ADCI includes several preset [[main:imageselectionmodel | image selection models]]. These image selection models are available for use directly after installation and do not have to be manually designed. Depending on which types of image quality issues are most prevalent in a sample, one of the preset models may be most appropriate to address those specific issues. This page will be updated with additional filters and filter descriptions which will aid the process of choosing the most appropriate filter. ADCI includes several preset [[main:imageselectionmodel | image selection models]]. These image selection models are available for use directly after installation and do not have to be manually designed. Depending on which types of image quality issues are most prevalent in a sample, one of the preset models may be most appropriate to address those specific issues. This page will be updated with additional filters and filter descriptions which will aid the process of choosing the most appropriate filter.
  
-===== Filter 1 (name pending) ===== +Preset image selection models are named according to the following convention: [X_Y] where X denotes a group of filtering methods and Y denotes a ranking type.
-  * Exclude if length-width ratio z-score > 1.5 +
-  * Exclude if centromere candidate density z-score > 1.5 +
-  * Exclude if finite difference z-score < -1.5 +
-  * Image quality ranking method: Group bin method +
-  * Include if ranked within the top 250 images+
  
-===== Filter 2 (name pending) ===== +===== Filtering method groups ===== 
-  * Exclude if length-width ratio z-score > 1.5 +**A**: No filtering methods\\  
-  * Exclude if centromere candidate density z-score > 1.5 +**B**: Object count filtering only\\  
-  * Exclude if finite difference z-score < -1.5 +**C**: Length-Width ratio, centromere candidate density, and finite difference filters\\  
-  * Exclude if object count < 40 +**D**: All six filtering methods 
-  * Exclude if object count > 60 + 
-  * Exclude if segmented count < 35 or > 50 +===== Ranking type =====  
-  * Exclude if classified object ratio > 0.6 +**A**: No image ranking\\  
-  * Image quality ranking method: None+**B**: Group bin method\\  
 +**C**: Combined filter score, custom filter weight 4,3,4,5,2,1\\  
 +**D**: Combined filter score, custom filter weight 5,2,4,3,4,1\\  
 +**E**: Combined filter score, custom filter weight 1,2,1,5,1,5\\
main/presetimgselectionmodels.txt · Last modified: 2017/05/16 16:24 by yli

Page Tools

Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki