CUPSEED - A combined use of prediction syntax elements to embed data in SHVC video

Area of Science:

• Digital video technology
• Information security
• Video compression standards

Background:

• Video data embedding is crucial for watermarking, steganography, and authentication.
• Conventional methods often use single syntax elements, leading to limitations in payload, quality, or bit rate.
• Scalable High Efficiency Video Coding (SHVC) offers potential for improved data embedding.

Purpose of the Study:

• To explore the combined use of multiple prediction syntax elements in SHVC for enhanced data embedding.
• To overcome the limitations of conventional single-syntax-element data embedding methods.
• To achieve higher payload capacity and maintain perceptual video quality.

Main Methods:

• Manipulating multiple prediction syntax elements within SHVC: intra prediction mode, motion vector predictor, motion vector difference, merge mode, and coding block structure.
• Comparing the proposed multi-element embedding approach against conventional single-element methods.
• Evaluating payload capacity, perceptual quality (PSNR), and bit rate overhead.

Main Results:

• The combined use of prediction syntax elements significantly increases payload capacity compared to single-element methods.
• Perceptual video quality is preserved with minimal degradation.
• In the 'PartyScene' sequence, 556.1 kbps was embedded with only a 0.15 dB PSNR drop and 7.4% bit rate overhead.
• Minimal bit rate variation was observed.

Conclusions:

• Combining multiple prediction syntax elements in SHVC is an effective strategy for high-payload, high-quality video data embedding.
• The proposed method outperforms conventional approaches in terms of payload and quality preservation.
• Recommendations are provided for selecting specific syntax elements based on video characteristics for optimal data embedding.