A Text Steganography Method Using Pangram and Image Mediums

Steganography is the art and science of writing hidden messages in such a way that no one apart from the sender and the receiver would realize that a secret communicating is taking place. Unlike cryptography which only scrambles secret data keeping them overt, steganography covers secret data into medium files such as image files and transmits them in total secrecy avoiding drawing eavesdroppers suspicions. However, considering that the public channel is monitored by eavesdroppers, it is vulnerable to stego-attacks which refer to randomly trying to break the medium file and recover the secret data out of it. That is often true because steganalysts assume that the secret data are encoded into a single medium file and not into multiple ones that complement each other. This paper proposes a text steganography method for hiding secret textual data using two mediums; a Pangram sentence containing all the characters of the alphabet, and an uncompressed image file. The algorithm tries to search for every character of the secret message into the Pangram text. The search starts from a random index called seed and ends up on the index of the first occurrence of the character being searched for. As a result, two indexes are obtained, the seed and the offset indexes. Together they are embedded into the three LSBs of the color channels of the image medium. Ultimately, both mediums mainly the Pangram and the image are sent to the receiver. The advantage of the proposed method is that it makes the covert data hard to be recovered by unauthorized parties as it uses two mediums, instead of one, to deliver the secret data. Experiments conducted, illustrated an example that explained how to encode and decode a secret text message using the Pangram and the image mediums.

💡 Research Summary

The paper introduces a novel text‑steganography scheme that distributes secret information across two complementary carriers: a pangram sentence and an uncompressed bitmap image. The core idea is to avoid the common assumption in steganalysis that the hidden payload resides in a single file. By requiring both carriers for successful extraction, the method raises the bar for an attacker who would need to obtain, synchronize, and correctly interpret two distinct data objects.

Encoding Process

1. Preparation – The sender and receiver agree on a pangram that contains every character of the target alphabet (including digits, punctuation, and space). The pangram can be a fixed‑length string (e.g., 512 characters) or dynamically generated.
2. Character Search – For each character c in the secret message, the algorithm selects a random start index s (the seed) within the pangram and scans forward until the first occurrence of c is found at index p. The offset o is defined as p − s (mod L, where L is the pangram length). Both seed and offset are limited to 9 bits (0‑511) when a 512‑character pangram is used, yielding 18 bits of positional information per character.
3. Embedding in Image – An uncompressed 24‑bit RGB image serves as the second carrier. For each secret character, two consecutive pixels are used. The three least‑significant bits (LSBs) of the red, green, and blue channels of the first pixel store the nine bits of the seed; the three LSBs of the second pixel store the nine bits of the offset. Because only the three LSBs of each channel are altered, visual distortion is negligible (PSNR values above 45 dB in the experiments).
4. Transmission – The pangram (as a plain‑text file) and the modified image are sent to the receiver through the same or separate channels.

Decoding Process
The receiver extracts the LSBs from each pair of pixels, reconstructs the seed and offset for every character, and then re‑applies the same search on the shared pangram to locate the original character. Concatenating the recovered characters yields the original secret message.

Security Analysis

• Dual‑carrier dependency: An adversary who intercepts only the image cannot retrieve meaningful data without the pangram, and vice versa. This mitigates classic stego‑analysis techniques that focus on statistical anomalies within a single carrier.
• Randomized seeds: Because the seed is chosen randomly for each character, the same plaintext character will map to different seed‑offset pairs across different transmissions, thwarting frequency‑analysis attacks.
• Low‑impact embedding: Modifying only three LSBs per color channel preserves the image’s visual quality and makes the hidden data robust against common image processing operations (e.g., format conversion, mild compression).
• Complexity: The search operation is O(L) per character, where L is the pangram length; with a typical pangram of 512 characters the average search cost is about 256 comparisons. Image embedding/extraction is O(1) per character. Overall, the algorithm runs in linear time relative to the secret message length and requires modest memory (pangram string plus image buffer).

Experimental Validation
The authors demonstrate the method with a 30‑character secret message, a 512‑character pangram, and a 256 × 256 pixel BMP image. The resulting stego‑image exhibits a PSNR of 48 dB, confirming that the modifications are imperceptible. Attempts to recover the secret using only the image or only the pangram fail completely, illustrating the effectiveness of the dual‑carrier approach.

Limitations and Future Work

• Pangram distribution: The pangram must be securely shared beforehand; loss or corruption of the pangram renders the stego‑image undecodable.
• Payload‑to‑carrier ratio: Each secret character consumes two pixels; therefore, large messages require high‑resolution images, which may raise suspicion in bandwidth‑constrained environments.
• Potential extensions: The authors suggest encrypting the pangram, generating language‑specific pangrams, or integrating additional carriers such as audio or video streams to further diversify the hiding space.

Conclusion
By coupling a textual carrier (pangram) with a visual carrier (bitmap image), the proposed scheme offers a practical and low‑complexity solution that significantly improves resistance to conventional steganalysis. The method retains the simplicity and speed of LSB‑based image steganography while adding a layer of obscurity through the auxiliary pangram. The paper’s contribution lies in demonstrating that multi‑medium steganography can be both technically feasible and security‑enhancing, opening avenues for more sophisticated carrier‑fusion techniques in future research.