[algorithm] fixed the import to convert and prepare feature, and the single score for one sequence

bob/pad/base/algorithm/OneClassGMM.py

@@ -18,7 +18,7 @@ import bob.io.base
 
 from sklearn import mixture
 
-from bob.pad.base.utils import convert_frame_cont_to_array, mean_std_normalize
+from bob.pad.base.utils import convert_frame_cont_to_array, mean_std_normalize, convert_and_prepare_features
 
 # ==============================================================================
 # Main body :
@@ -184,7 +184,7 @@ class OneClassGMM(Algorithm):
         """
 
         # training_features[0] - training features for the REAL class.
-        real = self.convert_and_prepare_features(
+        real = convert_and_prepare_features(
             training_features[0])  # output is array
 
         # training_features[1] - training features for the ATTACK class.
@@ -347,6 +347,8 @@ class OneClassGMM(Algorithm):
             If ``frame_level_scores_flag = True`` a list of scores is returned.
             One score per frame/sample.
         """
+        if toscore.ndim == 1:
+          return [toscore[0]]
 
         if self.frame_level_scores_flag: