Tamper Localization of the Medical Images Based on Fragile Watermarking

ÜSTÜBİOĞLU A., ULUTAŞ G., ÜSTÜBİOĞLU B.

25th Signal Processing and Communications Applications Conference (SIU), Antalya, Türkiye, 15 - 18 Mayıs 2017 identifier identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Cilt numarası:
  • Doi Numarası: 10.1109/siu.2017.7960481
  • Basıldığı Şehir: Antalya
  • Basıldığı Ülke: Türkiye
  • Karadeniz Teknik Üniversitesi Adresli: Evet

Özet

Flexible usage and easy acquisition of image processing software has made "tamper localization in medical images" a popular area. Medical Image Watermarking (MIW), which is one of the recommended methods in this area, relies on embedding additional information into the original image. In general, the medical image is separated into two parts called ROI - Region of Interest and RONI - Region of NonInterest. In the proposed method, image is separated into non-overlapping 8x8 blocks and then the BTLV (Block Tamper Localization Vector) of the corresponding block is calculated and embedded to tamper localization on the ROI. Similar works in the literature determined small size in ROI but it can reach up %80 of the image in the proposed method. Experimental results also indicate that proposed method gives higher PSNR for watermarked images and Normalize Correlation (NC) value is also better than the similar works.