Published on Wed Dec 16 2020

Self-Supervised Person Detection in 2D Range Data using a Calibrated Camera

Dan Jia, Mats Steinweg, Alexander Hermans, Bastian Leibe

Deep learning is the essential building block of state-of-the-art person-detectors in 2D range data. Only a few annotated datasets are currently available for training and testing these deep networks. We propose a method, which uses bounding boxes from an image-based.

4
0
0
Abstract

Deep learning is the essential building block of state-of-the-art person detectors in 2D range data. However, only a few annotated datasets are available for training and testing these deep networks, potentially limiting their performance when deployed in new environments or with different LiDAR models. We propose a method, which uses bounding boxes from an image-based detector (e.g. Faster R-CNN) on a calibrated camera to automatically generate training labels (called pseudo-labels) for 2D LiDAR-based person detectors. Through experiments on the JackRabbot dataset with two detector models, DROW3 and DR-SPAAM, we show that self-supervised detectors, trained or fine-tuned with pseudo-labels, outperform detectors trained only on a different dataset. Combined with robust training techniques, the self-supervised detectors reach a performance close to the ones trained using manual annotations of the target dataset. Our method is an effective way to improve person detectors during deployment without any additional labeling effort, and we release our source code to support relevant robotic applications.

Mon Jun 21 2021
Computer Vision
Domain and Modality Gaps for LiDAR-based Person Detection on Mobile Robots
Person detection is a crucial task for mobile robots navigating in human-populated environments. LiDAR sensors are promising for this task, given their accurate depth measurements and large field of view. We conduct a series of experiments, using the recently released JackRabbot dataset and the state-
2
0
0
Fri Apr 06 2018
Computer Vision
Deep Person Detection in 2D Range Data
Detecting humans is a key skill for mobile robots and intelligent vehicles. Detecting people from 2D range data is challenging due to the speed and dynamics of human leg motion. We recently developed the deep learning based wheelchair and walker detector DROW.
0
0
0
Fri Aug 13 2021our pick
Computer Vision
Is Pseudo-Lidar needed for Monocular 3D Object detection?
DD3D is an end-to-end, single stage, monocular 3D object detector. It can benefit from depth pre-training like pseudo-lidar methods, but without limitations.
0
0
0
Tue Mar 08 2016
Neural Networks
DROW: Real-Time Deep Learning based Wheelchair Detection in 2D Range Data
We introduce the DROW detector, a deep learning based detector for 2D range data. Laser scanners are lighting invariant, provide accurate range data, and typically cover a large field of view. So far, research on detection in laser range data has been dominated by hand-crafted features.
0
0
0
Wed Apr 29 2020
Computer Vision
DR-SPAAM: A Spatial-Attention and Auto-regressive Model for Person Detection in 2D Range Data
0
0
0
Tue Jan 29 2019
Machine Learning
Deep Active Learning for Efficient Training of a LiDAR 3D Object Detector
Training a deep object detector for autonomous driving requires a huge amount of labeled data. Active learning has the potential to minimize human annotation efforts while maximizing the object detector's performance.
0
0
0
Wed Nov 20 2019
Computer Vision
EfficientDet: Scalable and Efficient Object Detection
The EfficientDet-D7 achieves state-of-the-art 55.1 AP on COCO test-dev with 77M parameters and 410B FLOPs. It is 4x - 9x smaller and using 13x - 42x fewer FLOP's than previous
4
3
20
Mon Jun 08 2015
Computer Vision
You Only Look Once: Unified, Real-Time Object Detection
YOLO is a new approach to object detection. A single neural network predicts bounding boxes and class probabilities directly from full images. YOLO makes more localization errors but is far less likely to predict false detections where nothing exists.
4
2
6
Wed Oct 25 2017
Machine Learning
mixup: Beyond Empirical Risk Minimization
Large deep neural networks are powerful, but exhibit undesirable behaviors such as memorization and sensitivity to adversarial examples. In this work, we propose mixup, a simple learning principle to alleviate these issues. In essence, mixup trains a neural network on convex combinations of pairs of examples and
3
1
4
Fri Aug 16 2019
Machine Learning
Symmetric Cross Entropy for Robust Learning with Noisy Labels
DNN learning with Cross Entropy (CE) exhibits overfitting to noisy labels on some classes. CE requires an extra term to facilitate learning of hard classes. Inspired by the symmetric KL-divergence, we propose the approach of Symmetric cross entropy Learning (SL)
1
0
3
Mon Dec 22 2014
Machine Learning
Adam: A Method for Stochastic Optimization
Adam is an algorithm for first-order gradient-based optimization of stochastic objective functions. The method is straightforward to implement and has little memory requirements. It is well suited for problems that are large in terms of data and parameters.
3
0
2
Tue Dec 08 2015
Computer Vision
SSD: Single Shot MultiBox Detector
We present a method for detecting objects in images using a single deep neurological network. Our approach, named SSD, discretizes the output space of bounding boxes into a set of default boxes over different aspect ratios. At prediction time, the network generates scores for the presence of each object category in each default box.
1
0
1