Linear Span Network for Object Skeleton Detection

TL;DR

This paper introduces the Linear Span Network (LSN), a novel framework for object skeleton detection that enhances feature fusion and independence, leading to improved accuracy in reconstructing object skeletons.

Contribution

The paper proposes the Linear Span framework and LSN with Linear Span Units, which optimize feature reconstruction and independence for better skeleton detection.

Findings

Achieves state-of-the-art performance in object skeleton detection.

Effectively suppresses background clutter and reconstructs skeletons.

Enhances feature independence and fusion efficiency.

Abstract

Robust object skeleton detection requires to explore rich representative visual features and effective feature fusion strategies. In this paper, we first re-visit the implementation of HED, the essential principle of which can be ideally described with a linear reconstruction model. Hinted by this, we formalize a Linear Span framework, and propose Linear Span Network (LSN) modified by Linear Span Units (LSUs), which minimize the reconstruction error of convolutional network. LSN further utilizes subspace linear span beside the feature linear span to increase the independence of convolutional features and the efficiency of feature integration, which enlarges the capability of fitting complex ground-truth. As a result, LSN can effectively suppress the cluttered backgrounds and reconstruct object skeletons. Experimental results validate the state-of-the-art performance of the proposed LSN.