Method

Road segmentation using reverse and boundary attention [RBANet]

Submitted on 22 Apr. 2019 11:49 by
Yewon Kim (Korea Univ.)

 Running time: 0.16 s Environment: GPU @ 1.5 Ghz (Python + C/C++)

 Method Description: Segnet + Reverse attention + boundary attention Parameters: 22M Latex Bibtex: @inproceedings{sun2019reverse, title={Reverse and Boundary Attention Network for Road Segmentation}, author={Sun, Jee-Young and Kim, Seung-Wook and Lee, Sang-Won and Kim, Ye-Won and Ko, Sung-Jea}, booktitle={Proceedings of the IEEE International Conference on Computer Vision Workshops}, pages={0--0}, year={2019} }

Evaluation in Bird's Eye View

 Benchmark MaxF AP PRE REC FPR FNR UM_ROAD 95.78 % 89.14 % 94.92 % 96.66 % 2.36 % 3.34 % UMM_ROAD 97.38 % 92.67 % 96.70 % 98.08 % 3.68 % 1.92 % UU_ROAD 94.91 % 86.35 % 92.53 % 97.42 % 2.56 % 2.58 % URBAN_ROAD 96.30 % 89.72 % 95.14 % 97.50 % 2.75 % 2.50 %
This table as LaTeX

Behavior Evaluation

 Benchmark PRE-20 F1-20 HR-20 PRE-30 F1-30 HR-30 PRE-40 F1-40 HR-40
This table as LaTeX

The following plots show precision/recall curves for the bird's eye view evaluation.

Distance-dependent Behavior Evaluation

The following plots show the F1 score/Precision/Hitrate with respect to the longitudinal distance which has been used for evaluation.

Visualization of Results

The following images illustrate the performance of the method qualitatively on a couple of test images. We first show results in the perspective image, followed by evaluation in bird's eye view. Here, red denotes false negatives, blue areas correspond to false positives and green represents true positives.

 This figure as: png This figure as: png This figure as: png This figure as: png This figure as: png This figure as: png This figure as: png This figure as: png This figure as: png This figure as: png This figure as: png This figure as: png This figure as: png This figure as: png This figure as: png This figure as: png This figure as: png This figure as: png