The KITTI Vision Benchmark Suite

Method

structured random forest [SRF]

Submitted on 1 Apr. 2016 05:01 by
Liang Xiao (National University of Defense Technology)

Running time:		0.2 s
Environment:		1 core @ 2.5 Ghz (C/C++)

Method Description:

Predict each patch with a structured label patch by
structured random forest.

Parameters:

10 trees with depth 20
patchsize = 24

Latex Bibtex:

@article{Xiao2016IJARS,
title={Monocular Road Detection Using Structured
Random Forest},
author={Xiao, Liang and Dai, Bin and Liu, Daxue
and Zhao, Dawei and Wu, Tao},
journal={Int J Adv Robot Syst},
volume={13},
pages={101},
year={2016}
}

Evaluation in Bird's Eye View

Benchmark	MaxF	AP	PRE	REC	FPR	FNR
UM_ROAD	76.43 %	83.24 %	75.53 %	77.35 %	11.42 %	22.65 %
UMM_ROAD	90.77 %	92.44 %	89.35 %	92.23 %	12.08 %	7.77 %
UU_ROAD	76.07 %	79.97 %	71.47 %	81.31 %	10.57 %	18.69 %
URBAN_ROAD	82.44 %	87.37 %	80.60 %	84.36 %	11.18 %	15.64 %

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

Road/Lane Detection

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

This figure as: png eps pdf

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