Method

V2FedR: A Federated Framework for Driveable Road Area Detection [V2FedR]
[Anonymous Submission]

Submitted on 12 Sep. 2024 10:13 by
[Anonymous Submission]

Running time:0.05 s
Environment:GPU @ >3.5 Ghz (Python)

Method Description:
Privacy-Preserving Road Segmentation in Autonomous
Driving with Federated Vision
Parameters:
None
Latex Bibtex:

Evaluation in Bird's Eye View


Benchmark MaxF AP PRE REC FPR FNR
UM_ROAD 95.21 % 88.30 % 94.00 % 96.46 % 2.81 % 3.54 %
UMM_ROAD 96.61 % 91.12 % 94.99 % 98.29 % 5.70 % 1.71 %
UU_ROAD 94.07 % 86.50 % 92.69 % 95.49 % 2.45 % 4.51 %
URBAN_ROAD 95.58 % 88.81 % 94.14 % 97.07 % 3.33 % 2.93 %
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.


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