A tamper detection scheme based on integer wavelet transform for authenticating image
Advanced technology allows users to alter image and distribute the tampered image to social media. This research proposes a set level of tamper detection for identifying the integrity of image. The image is transformed by using integer wavelet transform (IWT), the authentication bits are generated f...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | en |
| Published: |
Elsevier
2025
|
| Subjects: | |
| Online Access: | https://umpir.ump.edu.my/id/eprint/46208/1/_A%20tamper%20detection%20scheme%20based%20on%20integer%20wavelet%20transform%20for%20authenticating%20image.pdf https://10.1016/j.fraope.2025.100344 https://umpir.ump.edu.my/id/eprint/46208/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Advanced technology allows users to alter image and distribute the tampered image to social media. This research proposes a set level of tamper detection for identifying the integrity of image. The image is transformed by using integer wavelet transform (IWT), the authentication bits are generated from bitwise XOR from approximation coefficients (ACb) sub-band and random binary matrix. The authentication bits are then embedded into the horizontal (HCb) and vertical (VCb) sub-bands. Cross-detect authentication bits of the HCb and VCb are used for detecting the tampered image. The proposed set level of tamper detection adopts the tic-tac�toe theorem to enhance the detection accuracy. The experimental results show that our proposed set level of tamper detection has improved the accuracy and precision of detection against malicious and incidental attacks. The proposed scheme achieved PSNR value of 54.73 dB and an SSIM value of 0.9987 of the watermarked images. In tamper detection, the proposed set level of tamper detection obtained an average accuracy of 0.9705 and a recall of 0.9443 under various attacks. Furthermore, the proposed scheme achieved a high accuracy value >0.97 against collage attack, copy-paste attack, and object deletion attack. |
|---|