Using Descriptor Counts in Clustering

Assumptions and Limitations

• Binary descriptors.
  
  
• Large dataset.
  
  
• Clusterable dataset, i.e.,
  • Not all singletons.
  • Not all the same (or nearly the same).