MCPNS: A Macropixel Collocated Position and Its Neighbors Search for Plenoptic 2.0 Video Coding

TL;DR

This paper introduces a specialized motion estimation algorithm for plenoptic 2.0 video coding, leveraging statistical analysis and a novel search pattern to improve efficiency and bitrate savings over existing methods.

Contribution

It presents a new fast motion estimation algorithm tailored for plenoptic 2.0 videos, utilizing macropixel collocated positions and neighbor searches based on motion characteristics.

Findings

Achieves better bitrate savings than existing methods.

Reduces computational complexity in motion estimation.

Effectively handles large motion deviations in plenoptic 2.0 sequences.

Abstract

Plenoptic 2.0 cameras enable high-resolution light field capture by incorporating focused optical designs that differ fundamentally from traditional plenoptic 1.0 systems. These structural differences produce distinct motion characteristics that challenge existing motion estimation (ME) algorithms. In this paper, we first conduct a comprehensive statistical analysis on real captured datasets to identify the primary differences in motion vector distributions among conventional, plenoptic 1.0, and plenoptic 2.0 videos. Building on these observations, we propose a novel fast ME algorithm specifically designed for plenoptic 2.0 video coding. The proposed method performs a joint search over macropixel collocated positions (MCPs) and their neighboring regions to effectively handle the large motion deviations typically observed in plenoptic 2.0 sequences. To further improve efficiency, we introduce a macropixel-level diamond search pattern (MLDSP) that follows the center-biased motion-vector distribution at the macropixel resolution, along with a fast MCP neighbor search restricted to the top K number of MCPs with the lowest distortion costs. Experimental results demonstrate that the proposed algorithm achieves better bitrate savings and computational complexity reductions compared to existing ME methods.