Bags of features for classification of Laser Scanning Microscopy data

The Bag-of-Features (BoF) paradigm has been introduced to the field of computer vision about a decade ago. Since then, its potential for image classification and retrieval tasks has been demonstrated in numerous experiments, which contributed to BoF approaches becoming well-established in the field. The BoF methods reported to date use mainly spatial intensity information but data collected by Laser Scanning Microscopy (LSM) techniques many times embed additional information that could be exploited in parallel in sophisticated BoF scenarios. In this contribution we discuss complementary LSM information categories that BoF frameworks could take advantage of when addressing the classification of LSM datasets.