An efficient post-processing adaptive filtering technique to rectifying the flickering effects.

Area of Science:

• Digital Image Processing
• Video Compression
• Signal Processing

Background:

• Low-bitrate video compression (≤0.5bpp) using standards like H.264 and MPEG-4 introduces significant visual artifacts.
• Common artifacts include quantization error, ringing, blocking, and flickering, degrading overall video quality.

Purpose of the Study:

• To develop an efficient pre- and post-processing technique (PP-AFT) for rectifying visual degradation in low-bitrate compressed video frames.
• To significantly improve the subjective visual quality of video frames affected by compression artifacts.

Main Methods:

• The PP-AFT method classifies video frames into regions using an activity function.
• Adaptive filters are designed for each classified region to address specific artifacts.
• Includes adaptive flicker extraction/removal and a 2-D filter for edge ringing reduction.

Main Results:

• PP-AFT demonstrated significant improvements in subjective video frame quality across various test videos.
• Outperformed state-of-the-art de-blocking methods, achieving 0.7-1.9db higher PSNR-B.
• Reduced average GBIM by 35.83-47.7% while maintaining MSSIM values close to the original (0.978).

Conclusions:

• The proposed PP-AFT technique effectively mitigates common video compression artifacts at very low bitrates.
• PP-AFT offers a superior solution for enhancing video quality compared to existing de-blocking methods.
• The method achieves a strong balance between artifact reduction and preservation of original video sequence fidelity.