Method

MSL3D: 3D Object Detection from Monocular, Stereo and Point Cloud for Autonomous Driving [MSL3D]
[Anonymous Submission]

Submitted on 13 Nov. 2020 03:44 by
[Anonymous Submission]

Running time:0.03 s
Environment:GPU @ 2.5 Ghz (Python)

Method Description:
In this paper, we propose a novel deep
architecture by combining multiple sensors for 3D
object detection, named MSL3D. While recently
LiDAR-Camera methods introduce additional semantic
cues, working with fewer false detections, there
is still a performance gap compare LiDAR-only
method. We argue that this gap is caused for two
reasons: 1) the 3D spherical receptive fields of
the set abstraction of the point clouds are not
aligned with the 2D pixel-level receptive fields
of the image. 2) the premature introduction of
image information makes it is difficult to apply
data augmentation both LiDAR and image
synchronously. For the first problem, we extend 3D
set abstraction to a 2D set abstraction that can
transform the 2D image features to the 3D sphere
to unify the receptive field of multi-modal data.
For the second problem, we design a novel two-
stage 3D detection framework that employs the
LiDAR-only backbone in the first stage to estimate
high-recall and high-quality proposals and then
Parameters:
no
Latex Bibtex:
no

Detailed Results

Object detection and orientation estimation results. Results for object detection are given in terms of average precision (AP) and results for joint object detection and orientation estimation are provided in terms of average orientation similarity (AOS).


Benchmark Easy Moderate Hard
Car (Detection) 95.76 % 94.60 % 92.16 %
Car (Orientation) 95.74 % 94.46 % 91.94 %
Car (3D Detection) 87.27 % 81.15 % 76.56 %
Car (Bird's Eye View) 91.64 % 88.23 % 85.53 %
Pedestrian (Detection) 69.07 % 58.57 % 55.86 %
Pedestrian (Orientation) 63.54 % 52.49 % 49.53 %
Pedestrian (3D Detection) 45.00 % 38.58 % 35.72 %
Pedestrian (Bird's Eye View) 48.81 % 42.82 % 40.13 %
Cyclist (Detection) 85.93 % 76.96 % 70.41 %
Cyclist (Orientation) 84.91 % 75.82 % 69.33 %
Cyclist (3D Detection) 76.74 % 62.27 % 56.20 %
Cyclist (Bird's Eye View) 81.23 % 68.57 % 62.01 %
This table as LaTeX


2D object detection results.
This figure as: png eps pdf txt gnuplot



Orientation estimation results.
This figure as: png eps pdf txt gnuplot



3D object detection results.
This figure as: png eps pdf txt gnuplot



Bird's eye view results.
This figure as: png eps pdf txt gnuplot



2D object detection results.
This figure as: png eps pdf txt gnuplot



Orientation estimation results.
This figure as: png eps pdf txt gnuplot



3D object detection results.
This figure as: png eps pdf txt gnuplot



Bird's eye view results.
This figure as: png eps pdf txt gnuplot



2D object detection results.
This figure as: png eps pdf txt gnuplot



Orientation estimation results.
This figure as: png eps pdf txt gnuplot



3D object detection results.
This figure as: png eps pdf txt gnuplot



Bird's eye view results.
This figure as: png eps pdf txt gnuplot




eXTReMe Tracker