Improving computational complexity using I-zag for text image authentication scheme
With the global proliferation of information and communication technologies, digital documents are increasingly used across various sectors like education, military, medical, business, and more. While the ease of access today brings convenience, it also enables easier data manipulation. Thus, authen...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2024
|
| Subjects: | |
| Online Access: | http://umpir.ump.edu.my/id/eprint/44083/ http://umpir.ump.edu.my/id/eprint/44083/1/Improving%20computational%20complexity%20using%20I-zag%20for%20text%20image%20authentication%20scheme.pdf |
| Summary: | With the global proliferation of information and communication technologies, digital documents are increasingly used across various sectors like education, military, medical, business, and more. While the ease of access today brings convenience, it also enables easier data manipulation. Thus, authentication mechanisms are imperative to secure sensitive user data, especially confidential text images. Most existing authentication schemes for text images are limited to specific image sizes and do not focus on processing time. As processing time is crucial for authentication systems, the primary objective of this thesis is to develop a scheme that improves the computational complexity of text image authentication using a fragile watermarking domain. The proposed scheme incorporates detection, localization, and recovery of tampered areas. A key contribution is utilizing Zigzag SCAN patterns for numbering and mapping prior to watermark embedding, along with Least Significant Bits (LSB) to embed the watermark bits. The authentication watermark bits are generated via average intensity comparison and parity check bits on the original cover image. This embedding approach facilitates hierarchical tamper detection and recovery. The proposed I-ZAG technique yields the fastest processing time and most efficient watermarking for diverse text image types, sizes, and sources. Experiments show I-ZAG averages 28.62 seconds for embedding across formats, with high 58 dB average PSNR and low 0.087% MSE. This surpasses existing schemes, Spiral-LSB and Hilbert-LSB. Detection, localization, and recovery are also fast. I-ZAG achieves ~100% localization accuracy and recovery rates. By enhancing computational complexity, I-ZAG enables rapid processing alongside robust authentication suitable for any text image. Significantly, I-ZAG meets the research objectives and addresses the need for an efficient digital watermarking scheme with low complexity to authenticate confidential text images. |
|---|