An Enhanced Texture-Based Feature Extraction Approach for Classification of Biomedical Images of CT-Scan of Lungs

• Varun Srivastava ^[3] ; Shilpa Gupta ^[3] ; Gopal Chaudhary ^[3] ; Arun Balodi ^[4] ; Manju Khari ^[1] ; Vicente García-Díaz ^[2]
  1. [1] University of Technology

     University of Technology

     Rusia

     
  2. [2] Universidad de Oviedo

     Universidad de Oviedo

     Oviedo, España

     
  3. [3] Bharati Vidyapeeth’s College of Engineering, Paschim Vihar, New Delhi
  4. [4] Atria Institute of Technology, Bengaluru, Karnataka

  Mostrar afiliaciones +
• Localización: IJIMAI, ISSN-e 1989-1660, Vol. 6, Nº. 7, 2021, págs. 18-25
• Idioma: inglés
• Enlaces
  • Texto completo (pdf)
• Resumen

  • Content Based Image Retrieval (CBIR) techniques based on texture have gained a lot of popularity in recent times. In the proposed work, a feature vector is obtained by concatenation of features extracted from local mesh peak valley edge pattern (LMePVEP) technique; a dynamic threshold based local mesh ternary pattern technique and texture of the image in five different directions. The concatenated feature vector is then used to classify images of two datasets viz. Emphysema dataset and Early Lung Cancer Action Program (ELCAP) lung database. The proposed framework has improved the accuracy by 12.56%, 9.71% and 7.01% in average for data set 1 and 9.37%, 8.99% and 7.63% in average for dataset 2 over three popular algorithms used for image retrieval.