Automatic Identification of Ficus deltoidea Jack (Moraceae) Varieties Based on Leaf


  •  A Fakhri Nasir    
  •  M Nordin A Rahman    
  •  Nashriyah Mat    
  •  Rasid Mamat    

Abstract

Currently, the traditional method used to identify Ficus deltoidea Jack (Moraceae) varieties require the plant taxonomists to observe and examine the leaf morphology of herbarium or live specimens. An automated variety identification system would ease the herbs collector to carry out valuable plant identification work. In this paper, a model for F. deltoidea varieties identification based on their leaf shape, color and texture was developed. Five different varieties of F. deltoidea were used in the proposed work with sixty nine sample data collected for each of varieties. First, the F. deltoidea leaves were plucked and the picture of leaves is then taken by a digital scanner in the format of JPEG.  For leaf shape, a total of fourteen shape features were extracted based on basic geometric features. The mean of different color channels was calculated in leaf color feature extraction. Furthermore, four texture features based on gray-level co-occurrence matrix was implemented to extract leaf texture properties. By using the leaf structure, a set of three different leaf properties which are leaf shape, color and texture features was extracted. The features weight is then calculated using eigenvalues coefficient in principal component analysis. The best principal components are retained for identification experiments. Lastly, Nearest Neighbor with Euclidean distance was used in variety identification based on three different leaf properties mentioned above. The effectiveness of different leaf features are demonstrated in the identification experiment.

 



This work is licensed under a Creative Commons Attribution 4.0 License.
  • Issn(Print): 1913-1844
  • Issn(Onlne): 1913-1852
  • Started: 2007
  • Frequency: monthly

Journal Metrics

(The data was calculated based on Google Scholar Citations)

Google-based Impact Factor (2018): 6.49

h-index (January 2018): 30

i10-index (January 2018): 163

h5-index (January 2018): 19

h5-median(January 2018): 25

Contact