Line detection via a lightweight CNN with a Hough Layer

TL;DR

This paper introduces a lightweight CNN with an embedded Hough layer for improved line detection, addressing limitations of traditional CNNs and dataset inconsistencies.

Contribution

It proposes a novel CNN architecture with a parameter-free Hough layer that enhances line detection capabilities.

Findings

The Hough layer enables global receptive fields for better line detection.

The approach outperforms traditional CNNs on benchmark datasets.

Identifies dataset inconsistencies affecting line detection evaluation.

Abstract

Line detection is an important computer vision task traditionally solved by Hough Transform. With the advance of deep learning, however, trainable approaches to line detection became popular. In this paper we propose a lightweight CNN for line detection with an embedded parameter-free Hough layer, which allows the network neurons to have global strip-like receptive fields. We argue that traditional convolutional networks have two inherent problems when applied to the task of line detection and show how insertion of a Hough layer into the network solves them. Additionally, we point out some major inconsistencies in the current datasets used for line detection.