Improving Spatial Codification in Semantic Segmentation

TL;DR

This paper introduces a new spatial codification method for semantic segmentation that partitions images into figure, border, and ground regions, and applies spatial pyramids to enhance object description, leading to improved segmentation results.

Contribution

It proposes a novel spatial partitioning and pyramid-based descriptor that reduces context influence and enhances object representation in semantic segmentation.

Findings

Improved segmentation accuracy on Pascal VOC datasets.

Enhanced object description with spatial pyramids.

Better Figure-Ground pooling performance.

Abstract

This paper explores novel approaches for improving the spatial codification for the pooling of local descriptors to solve the semantic segmentation problem. We propose to partition the image into three regions for each object to be described: Figure, Border and Ground. This partition aims at minimizing the influence of the image context on the object description and vice versa by introducing an intermediate zone around the object contour. Furthermore, we also propose a richer visual descriptor of the object by applying a Spatial Pyramid over the Figure region. Two novel Spatial Pyramid configurations are explored: Cartesian-based and crown-based Spatial Pyramids. We test these approaches with state-of-the-art techniques and show that they improve the Figure-Ground based pooling in the Pascal VOC 2011 and 2012 semantic segmentation challenges.