usage:feature_extraction

Differences

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

Link to this comparison view

Both sides previous revision Previous revision
Next revision		 Previous revision
usage:feature_extraction [2019/08/09 17:20] – [Outputs] pseudomoanerusage:feature_extraction [2020/06/26 15:33] (current) – [Choosing features to extract] pseudomoaner
Line 21: Line 21:
  
 Don't select features if you don't need them! This will reduce processing times and space requirements. Don't select features if you don't need them! This will reduce processing times and space requirements.
 +
 +</note>
 +
 +<note tip>
 +
 +If you want to measure a feature that you can't see here, you may be able to write a new function to extract it using FAST's integrated custom feature extraction system. [[usage:advanced_usage#custom_feature_extraction|Please see the documentation on this topic to see whether this could work for you.]]
  
 </note> </note>
Line 51: Line 57:
   * **trackableData:** The feature data associated with each object at each timepoint.   * **trackableData:** The feature data associated with each object at each timepoint.
  
-The trackableData structure forms the basis of the feature tracking module, and can be used in its own right for [[extra:faqs#what_is_image_cytometry_can_it_be_performed_by_fast|image cytometry]]. It is organised as shown in the below figure:+The trackableData structure forms the basis of the feature tracking module, and can be used by itself for [[extra:faqs#what_is_image_cytometry_can_it_be_performed_by_fast|image cytometry]]. It is organised as shown in the below figure: 
 + 
 +{{ :usage:trackabledatadepiction.png?nolink&700 |}} 
 + 
 +Values stored in identically indexed locations in the different fields of trackableData belong to the same object. For example, the width of object 10 at time 20 is stored in  
 + 
 +  trackableData.Width{20}(10) 
 + 
 +while its orientation is stored in  
 + 
 +  trackableData.Orientation{20}(10) 
 +   
 +For fields that contain multi-dimensional arrays rather than vectors (such as 'Centroid'), the object index corresponds to the row of the array. For example, the $x,y$-coordinates of the above object are stored in 
 + 
 +  trackableData.Centroid{20}(10,:) 
 + 
 +All geometrical measurements (length, width, area, position and orientation) are stored in physical units, while fluorescence-associated measurements (channel mean and channel standard deviation) are stored in arbitrary units.
  
 +===== Video demonstration =====
  
 +{{ youtube>h6kEowehxMo?large }}
  
  • usage/feature_extraction.1565367624.txt.gz
  • Last modified: 2019/08/09 17:20
  • by pseudomoaner