Reducing inverse quantization numbers in intra frame for video transcoding architectures

Area of Science:

• Digital Signal Processing
• Video Compression
• Computer Vision

Background:

• Conventional transcoding schemes often overlook the relationship between consecutive quantizer step sizes, impacting efficiency.
• Reducing computational complexity in video transcoding is crucial for real-time applications and resource-constrained devices.

Purpose of the Study:

• To propose a novel transcoding architecture for minimizing inverse quantization numbers.
• To enhance computational efficiency and objective performance in video transcoding.

Main Methods:

• Developed a transcoding architecture based on the modulus of the ratio of current and previous quantization parameters.
• Analyzed the quantized area of quantization parameters to identify computational redundancies.
• Implemented computer simulations to evaluate the proposed scheme against conventional approaches.

Main Results:

• The proposed architecture effectively reduces computing complexity by optimizing inverse quantization.
• Demonstrated superior performance, including improved peak signal-to-noise ratio (PSNR), compared to existing methods.
• Validated the scheme's effectiveness through objective analysis and computer simulations.

Conclusions:

• The novel transcoding architecture offers significant advantages in reducing computational complexity.
• The proposed method achieves better objective quality metrics, making it a viable alternative for efficient video transcoding.