Method

One For All [OFA]
[Anonymous Submission]

Submitted on 11 Nov. 2018 12:37 by
[Anonymous Submission]

Running time:0.04 s
Environment:GPU @ 1.5 Ghz (Python)

Method Description:
we propose a model named OFA-Net. OFA means “One
For All”, which is we can get all we want with just
one network. We next explain what “all we want”
refers to. “All” not only means simultaneous
detection and segmentation, but stands for exploring
the mutual effect between the two tasks.
Parameters:
learning rate =1e-4
weight decay = 5e-4
Adam optimizer
Latex Bibtex:

Evaluation in Bird's Eye View


Benchmark MaxF AP PRE REC FPR FNR
UM_ROAD 92.08 % 82.73 % 87.87 % 96.72 % 6.08 % 3.28 %
UMM_ROAD 95.43 % 89.10 % 92.78 % 98.24 % 8.41 % 1.76 %
UU_ROAD 92.62 % 83.12 % 88.97 % 96.58 % 3.90 % 3.42 %
URBAN_ROAD 93.74 % 85.37 % 90.36 % 97.38 % 5.72 % 2.62 %
This table as LaTeX

Behavior Evaluation


This table as LaTeX

Road/Lane Detection

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


eXTReMe Tracker