featuresLG

 Feature detection and description applied to .mat behavior data. 

 See RECOGNITION_DEMO / FEATURESSM for general steps of detection / description and
 differences between this function and FEATURESSM.
 
 INPUTS
   nsets           - number of sets
   cliptypes       - types of clips (cell of strings)
   par_stfeatures  - parameters for feature detection [see featuresLGdetect]
   cubdesc         - cuboid descriptor [see featuresLGdesc]
   ncuboids        - number of cuboids to grab per .mat file [see featuresLGpca]
   kpca            - number of dimensions to reduce data to [see featuresLGpca]

 OUTUPTS
   DATASETS    - array of structs, will have fields:
           .IDX        - length N vector of clip types
           .ncilps     - N: number of clips
           .cubcount   - length N vector of cuboids counts for each clip clip
           .subs       - length N cell vector of sets of locations of cuboids
           .desc       - length N cell vector of cuboid descriptors
   cubdesc         - output of featuresLGpca
   cuboids         - output of featuresLGpca
   
 See also FEATURESLGDETECT, FEATURESLGPCA, FEATURESLGDESC, FEATURESLGCONV