Skip to Main content Skip to Navigation
Conference papers

Exploiting Fully Convolutional Networks for Fast Road Detection

Abstract : Road detection is a crucial task in autonomous navigation systems, it is responsible for delimiting the road area and hence the free and valid space for maneuvers. In this paper, we consider the visual road detection problem where, given an image, the objective is to classify every of its pixels into road or non-road. We address this task by proposing a convolutional neural network architecture. We are especially interested in creating a model which takes advantage of a large contextual window while maintaining a fast inference. We achieve this by using a Network-in-Network (NiN) architecture and by converting the model into a fully convolutional network after training. Experiments were conducted to evaluated the effects of different contextual window sizes (the amount of contextual information) and also to evaluate the NiN aspect of the proposed architecture. Finally, we summited our results to the KITTI road detection benchmark achieving results in line with other state-of-the-art methods while maintaining real-time inference. The benchmark results also reveal that the inference time of our approach is unique at this level of accuracy, being two orders of magnitude faster than other methods with similar performance.
Complete list of metadatas

Cited literature [19 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01260697
Contributor : Vincent Fremont <>
Submitted on : Friday, February 1, 2019 - 5:38:06 PM
Last modification on : Monday, March 25, 2019 - 12:08:49 PM
Long-term archiving on: : Friday, May 3, 2019 - 1:57:25 AM

File

ICRA16_1263_FI.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01260697, version 1

Collections

Citation

Caio César Teodoro Mendes, Vincent Frémont, Denis Fernando Wolf. Exploiting Fully Convolutional Networks for Fast Road Detection. 2016 IEEE International Conference on Robotics and Automation (ICRA), May 2016, Stockholm, Sweden. ⟨hal-01260697⟩

Share

Metrics

Record views

630

Files downloads

147