Secure Image Indexing Using Speeded Up Robust Features (SURF) Key Points and SHAKE256 Hashing
| Main Author: | Mohammad Awad Alfawair, Khalid Alkaabneh |
|---|---|
| Format: | Article Journal |
| Bahasa: | ang |
| Terbitan: |
, 2021
|
| Subjects: | |
| Online Access: |
https://zenodo.org/record/5168797 |
Daftar Isi:
- Due to the rapid development of the Internet and storing data, several security attacks have taken place, causing deep fears to computer users. On the other hand, many computer users mainly depend on different techniques to protect their data and keep them safe and secure, including their private images, since they are important and sensitive. Some currently applied procedures do not provide enough protection for the privacy of the stored images. Consequently, researchers are exerting great effort to search for images in a more secure way using different codes. The primary purpose of this thesis is to solve the issue of data security to reduce the chance or possibility of losing these images. A procedure has been introduced to search for images more safely to be restored safely depending on the content of these images themselves. Also, this suggested technique helps provide more image privacy and the ability to search for these images more accurately. In addition, using the index will inevitably lead to more privacy for stored data. Based on my thesis, I recommend we use the SURF Algorithm to define the essential image points m, we use the OTP Algorithm to increase the security level, and finally, we apply the SHAKE256 algorithm for more security. Thus, the highly-secured search is provided, and precise control of reaching these images and restoring them is enabled. The approach that I recommend has been tested in this thesis using MATLAB Pack, which is very effective. Then, the findings of the test were tested using 100 images from the database of Stanford University. Those images were classified into four groups: animal images, people and faces, nature, and buildings and terrain. The results showed that the images that have more points are more secured and highly protected. Another criterion, about the number of zeroes and ones of the image, has been applied to the results to provide more security and protection. I also implemented two criteria to test the security level using the same findings. These two criteria are the number of zeroes and ones and the change in the image itself. As a result, I concluded that when the number of zeroes is equal to the number of ones, this shows a high level of image security. In conclusion, my research findings indicated that implementing the suggested approach to protect images is more effective than other previous researches.