Distortion-free embedding in the optic disk of retina fundus images using complex-valued neural network
One of the downside of the current watermarking system is the inevitable distortion caused by data embedding. Some algorithms have been modified to produce a minimal distortion in the host image. however such small alterations may not be acceptable for medical images which are use for diagnosing and...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IDOSI Publication
2011
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/18439/ http://irep.iium.edu.my/18439/ http://irep.iium.edu.my/18439/3/WASJ_published.pdf |
| Summary: | One of the downside of the current watermarking system is the inevitable distortion caused by data embedding. Some algorithms have been modified to produce a minimal distortion in the host image. however such small alterations may not be acceptable for medical images which are use for diagnosing and treatment of diseases. This paper presents a distortion-free method for embedding data in the optic nerves of retina images (fundus) accomplished by Complex-Valued Neural Networks. Features of the images are modeled in Fast Fourier Transform to obtain the complex values which serves as inputs to Complex-Valued Neural Networks, (CVNN). The performance of the proposed technique has been evaluated using ROC graphs and Image Fidelity Measure (IFM). Results obtained indicate that the proposed algorithm can perfectly embed watermark in the host image without any distortion or loss of information which make both host and the watermarked image perceptually indistinguishable. The novelty of the proposed algorithm is that it guarantees a complete recovery of the embedded data during watermark extraction provided the network is trained properly and correct weights are used. Hence it can improve correct diagnoses and therefore useful in authentication system. |
|---|