Correctly identifying neuronal subsets is critical to multiple downstream methods in several areas of neuroscience research. The hippocampal interneuron characterization technology has achieved rapid development in recent years. However, capturing true neuronal features for accurate interneuron characterization and segmentation has remained elusive. In the current study, a novel global preserving estimate algorithm is used to capture the non-linearity in the features of hippocampal interneurons after factor Algorithm. Our results provide evidence for the effective integration of the original linear and nonlinear neuronal features and achieves better characterization performance on multiple hippocampal interneuron databases through array matching.