Application of genetic algorithm to hexagon-based motion estimation

Area of Science:

• Digital Signal Processing
• Computer Vision
• Information Technology

Background:

• Growing demand for high-quality audio and video due to technological advancements.
• Motion estimation is crucial for video coding, reducing redundancy between frames.
• Existing fast algorithms for motion estimation have limitations based on monotonic error assumptions.

Purpose of the Study:

• To propose a new motion estimation technique for video coding.
• To combine the strengths of hexagon-based search and genetic algorithms.
• To enhance encoding speed and accuracy in video compression.

Main Methods:

• Integration of hexagon-based search algorithm with genetic algorithm for motion estimation.
• Development of a novel search scheme overcoming limitations of traditional fast algorithms.
• Experimental evaluation of the proposed method against existing techniques.

Main Results:

• The proposed method achieves motion estimation closer to full search accuracy.
• Experimental results demonstrate improved encoding speed and accuracy.
• The combined approach outperforms traditional fast algorithms in certain scenarios.

Conclusions:

• The novel technique effectively enhances motion estimation in video coding.
• Combining hexagon-based search and genetic algorithms offers a promising direction for efficient video compression.
• The proposed method provides a viable alternative for achieving high-quality video with reduced computational cost.