Overview: Main Fundamentals for Steganography

1. INTRODUCTION

• Steganography hides and transmits data in apparently innocuous carriers to conceal that the data exists, complementing rather than replacing cryptography.The introduction notes that encrypting a message before embedding it can add another layer of protection.Source: §1 · p. 1
• The introduction distinguishes the original “cover” from embedded data and describes imperceptible marks in digital media as a way to help prevent unauthorized copying.It also presents information hiding as a broader term for hiding information in redundant cover data.Source: §1 · p. 1
• The paper reviews earlier work and presents the authors’ analysis of fundamental steganography methods used to hide data in multimedia.The introduction links renewed interest in concealment to computers’ digital representation of data.Source: §1 · p. 2

2. STEGANOGRAPHY TYPES

• The paper identifies three types of steganography: pure, secret-key, and public-key.The section introduces each type in turn.Source: §2 · p. 2

2.1 Pure Steganography

• Pure steganography requires no prior exchange of secret information; its security depends entirely on secrecy, and extraction recovers the message embedded in a cover.It is represented using covers, messages, an embedding function, and an extraction function.Source: §2.1 · p. 2
• Pure steganography avoids sharing a stego-key, but its protocols provide no security if an attacker knows the embedding method.The paper says it is preferred in most applications because no stego-key must be shared.Source: §2.1 · p. 2

2.2 Secret Key Steganography

• In secret-key steganography, the sender embeds a message in a cover using a secret key, and the receiver uses that key to extract it.The system is described as similar to a symmetric cipher.Source: §2.2 · p. 2

2.3 Public Key Steganography

• Public-key steganography uses a public key for embedding and a private key to reconstruct the message; one described approach encrypts with the receiver’s public key before embedding.The sender embeds a random-looking encrypted message, which the receiver extracts and decrypts with the private key.Source: §2.3 · pp. 2–3

3. GENERAL STEGANOGRAPHY SYSTEM

• The sender’s algorithm combines a cover message with an embedded message to produce a stego message, which the receiver uses the reverse process to decode.The cover and stego messages must have the same data type, but the embedded message may have another type.Source: §3 · p. 3
• The embedding algorithm may use a stego key, which is additional secret data that may be needed for hiding and is usually needed, or related to what is needed, for extraction.The sender begins with a cover message as input to the stego-system.Source: §3 · p. 3
• The receiver reverses the embedding process to extract the embedded message from the stego message.The description allows for the same key or a related key to be used in extraction.Source: §3 · p. 3

4. CHARACTERIZATION OF STEGANOGRAPHY SYSTEMS

• The relative importance of features that characterize steganographic methods’ strengths and weaknesses depends on the application.The section names capacity, robustness, undetectability, invisibility, and security as characteristics.Source: §4 · p. 3

4.1 Capacity

• Capacity is the total number of bits hidden and successfully recovered by the stego system.Source: §4.1 · p. 3

4.2 Robustness

• Robustness is the ability of embedded data to remain intact through transformations such as filtering, random noise, scaling, rotation, and loose compression.Source: §4.2 · p. 3

4.3 Undetectable

• Undetectability requires the image with the embedded message to remain consistent with a model of the source images; it is affected by message size and cover-image format.For embedding in image noise, the method should avoid statistical changes to that noise.Source: §4.3 · p. 3

4.4 Invisibility (Perceptual Transparency)

• Perceptual invisibility means an average human cannot distinguish carriers with hidden information from those without it, without significant degradation or loss of cover quality.Source: §4.4 · p. 3

4.5 Security

• Security means embedded information is not removable after discovery by an attacker, and depends on the total information about the algorithm and secret key.Source: §4.5 · p. 3

5. CLASSIFICATION OF STEGANOGRAPHY TECHNIQUES

• Fig. 4 presents six technique categories grouped by embedding modifications, though exact classification can be difficult.Source: §5 · p. 4, Figure 4 · p. 4

5.1 Substitution Systems

• Substitution systems replace insignificant cover parts with message bits; extraction depends on knowing where those bits were embedded.Source: §5.1 · p. 4

5.1.1 Least Significant Bit Substitution

• The least significant bit method replaces selected cover elements’ LSBs with message bits, which the receiver extracts and assembles.Source: §5.1.1 · p. 4

5.1.2 Pseudorandom Permutation

• With a random-access cover, message bits can be distributed across cover bits in a random order, making their sequence less predictable to an attacker.Source: §5.1.2 · p. 4

5.1.3 Image Downgrading and Cover Channels

• Image downgrading swaps a cover image’s four least significant bits with a same-sized secret image’s four most significant bits; cover regions can instead store a bit by parity.Source: §5.1.3 · p. 4

5.1.5 Palette-Based Image

• Palette-based images can carry data in color-vector LSBs or palette ordering, but reordering the palette destroys messages encoded in its order.Source: §5.1.5 · p. 4

5.1.6 Quantization and Dithering

• Quantization and dithering encode a bit by replacing a quantized adjacent-pixel difference with the nearest difference associated with that bit, then decode using the difference and stego-key.Source: §5.1.6 · pp. 4–5

• Binary-image substitution is vulnerable to transmission errors, while unused or reversed storage space can hide data without perceptually degrading its carrier.Source: §5.1.6 · p. 5

• Substitution and modification methods are vulnerable to signal processing and even small changes from lossy compression can destroy the hidden information.Source: §5.2 · p. 5
• Discrete cosine and wavelet transforms are examples of embedding vehicles, with information distributed through transform coefficients to resist attacks while remaining hard to detect.Source: §5.2 · p. 5
• Transform-domain embedding modifies selected coefficients and is described as generally more robust than other methods because unaffected coefficients can remain intact after processing.Source: §5.2 · p. 5

• Spread spectrum transmits a signal across more bandwidth than the minimum needed, using a data-independent code and synchronized reception for recovery.Source: §5.3 · p. 5
• Direct-sequence spread spectrum modulates a pseudorandom signal with a chip-rate-spread secret signal and adds it to the cover; frequency hopping rapidly changes carrier frequency.Source: §5.3 · p. 5
• Spreading a message across the cover is described as difficult to perceive and difficult to remove, and these techniques are widely used in watermarking.Source: §5.3 · p. 5

• Unlike substitution, distortion decoding requires the original cover so the receiver can measure changes and reconstruct the sender’s modification sequence.Source: §5.4 · p. 5
• A sender chooses a sequence of cover modifications corresponding to the secret message, and the receiver uses the differences to recover that sequence.Source: §5.4 · p. 5
• Text-based distortion methods can modulate line or word positions, or hide data in extra spaces, tabs, and line breaks that browsers ignore.Source: §5.4 · pp. 5–6

5.5 Cover Generation Techniques

• Some applications generate a digital object solely to serve as a cover for secret communication,.This contrasts with adding secret information to a specific cover using an embedding algorithm.Source: §5.5 · p. 6
• These cover-generation applications differ from methods that add secret information to an existing cover,.Source: §5.5 · p. 6
• Table 1 is titled “Weaknesses of Steganography Techniques.”Source: Table 1 · p. 6

5.6 Statistical Steganography

• Statistical steganography schemes embed one bit of information in a digital carrier.Source: §5.6 · p. 6
• The cover is modified so statistical characteristics change significantly when a 1 is transmitted; otherwise, it remains unchanged,.Source: §5.6 · p. 6
• The receiver must distinguish unmodified covers from modified ones, with a secret bit inserted into its corresponding block,.The cover is divided into disjoint blocks, and a block is changed only when its bit is 1.Source: §5.6 · p. 6

6. CONCLUSION

• The overview discusses pure, secret-key, and public-key steganography types.Source: §6 · p. 6
• It explains the steganographic transmission system between sender and receiver.Source: §6 · p. 6
• It reviews features characterizing method strengths and weaknesses, and groups techniques into six categories while noting that exact classification is sometimes impossible.Source: §6 · p. 6