# Geodesic Distance Histogram Feature for Video Segmentation

## TL;DR

This paper introduces a geodesic-distance-based histogram feature that encodes global information to enhance video segmentation, demonstrating significant improvements when integrated into existing frameworks on benchmark datasets.

## Contribution

The paper presents a novel geodesic histogram feature with adaptive weights and spatial pyramids, improving video segmentation performance across multiple algorithms.

## Key findings

- Significantly improved segmentation accuracy on benchmark datasets.
- Effective integration of geodesic histogram into existing frameworks.
- Enhanced global information encoding for better segmentation results.

## Abstract

This paper proposes a geodesic-distance-based feature that encodes global information for improved video segmentation algorithms. The feature is a joint histogram of intensity and geodesic distances, where the geodesic distances are computed as the shortest paths between superpixels via their boundaries. We also incorporate adaptive voting weights and spatial pyramid configurations to include spatial information into the geodesic histogram feature and show that this further improves results. The feature is generic and can be used as part of various algorithms. In experiments, we test the geodesic histogram feature by incorporating it into two existing video segmentation frameworks. This leads to significantly better performance in 3D video segmentation benchmarks on two datasets.