Distantly Supervised Road Segmentation

TL;DR

This paper introduces a distantly supervised method for road segmentation that uses only image-level labels and large image databases to generate weak pixel-wise masks, achieving high performance with less annotation effort.

Contribution

The authors propose a novel approach leveraging distant supervision and large image datasets to train a road segmentation model with minimal annotation.

Findings

Achieves 93.8% of fully supervised performance on Cityscapes

Uses significantly less annotation effort than traditional methods

Demonstrates effectiveness of distant supervision for pixel-wise segmentation

Abstract

We present an approach for road segmentation that only requires image-level annotations at training time. We leverage distant supervision, which allows us to train our model using images that are different from the target domain. Using large publicly available image databases as distant supervisors, we develop a simple method to automatically generate weak pixel-wise road masks. These are used to iteratively train a fully convolutional neural network, which produces our final segmentation model. We evaluate our method on the Cityscapes dataset, where we compare it with a fully supervised approach. Further, we discuss the trade-off between annotation cost and performance. Overall, our distantly supervised approach achieves 93.8% of the performance of the fully supervised approach, while using orders of magnitude less annotation work.