Adaptive LiDAR Sampling and Depth Completion using Ensemble Variance

This work considers the problem of depth completion, with or without image data, where an algorithm may measure the depth of a prescribed limited number of pixels. The algorithmic challenge is to choose pixel positions strategically and dynamically to maximally reduce overall depth estimation error. This setting is realized in daytime or nighttime depth completion for autonomous vehicles with a programmable LiDAR. Our method uses an ensemble of predictors to define a sampling probability over pixels. This probability is proportional to the variance of the predictions of ensemble members, thus highlighting pixels that are difficult to predict. By additionally proceeding in several prediction phases, we effectively reduce redundant sampling of similar pixels. Our ensemble-based method may be implemented using any depth-completion learning algorithm, such as a state-of-the-art neural network, treated as a black box. In particular, we also present a simple and effective Random Forest-based algorithm, and similarly use its internal ensemble in our design. We conduct experiments on the KITTI dataset, using the neural network algorithm of Ma et al. and our Random Forest based learner for implementing our method. The accuracy of both implementations exceeds the state of the art. Compared with a random or grid sampling pattern, our method allows a reduction by a factor of 4-10 in the number of measurements required to attain the same accuracy.

[1]  Zhao Chen,et al.  Estimating Depth from RGB and Sparse Sensing , 2018, ECCV.

[2]  Jens H. Schmid,et al.  Subwavelength integrated photonics , 2018, Nature.

[3]  Steven Lake Waslander,et al.  In Defense of Classical Image Processing: Fast Depth Completion on the CPU , 2018, 2018 15th Conference on Computer and Robot Vision (CRV).

[4]  David A. Cohn,et al.  Active Learning with Statistical Models , 1996, NIPS.

[5]  Naif Alajlan,et al.  Active learning for spectroscopic data regression , 2012 .

[6]  Shan Liu,et al.  Integrating Active Learning and Contextually Guide for Semantic Labeling of LiDAR Point Cloud , 2018, 2018 10th IAPR Workshop on Pattern Recognition in Remote Sensing (PRRS).

[7]  Aggelos K. Katsaggelos,et al.  Adaptive Image Sampling Using Deep Learning and Its Application on X-Ray Fluorescence Image Reconstruction , 2018, IEEE Transactions on Multimedia.

[8]  Klaus C. J. Dietmayer,et al.  Deep Active Learning for Efficient Training of a LiDAR 3D Object Detector , 2019, 2019 IEEE Intelligent Vehicles Symposium (IV).

[9]  Ross D. King,et al.  Active Learning for Regression Based on Query by Committee , 2007, IDEAL.

[10]  Truong Q. Nguyen,et al.  Depth Reconstruction From Sparse Samples: Representation, Algorithm, and Sampling , 2014, IEEE Transactions on Image Processing.

[11]  Enrique del Castillo,et al.  Query-by-committee improvement with diversity and density in batch active learning , 2018, Inf. Sci..

[12]  H. Sebastian Seung,et al.  Query by committee , 1992, COLT '92.

[13]  Dongrui Wu,et al.  Pool-Based Sequential Active Learning for Regression , 2018, IEEE Transactions on Neural Networks and Learning Systems.

[14]  Michael R. Watts,et al.  Large-scale nanophotonic phased array , 2013, Nature.

[15]  In So Kweon,et al.  Depth Completion with Deep Geometry and Context Guidance , 2019, 2019 International Conference on Robotics and Automation (ICRA).

[16]  Luis Salgado,et al.  Efficient spatio-temporal hole filling strategy for Kinect depth maps , 2012, Electronic Imaging.

[17]  Qiao Wang,et al.  VirtualWorlds as Proxy for Multi-object Tracking Analysis , 2016, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[18]  Ruigang Yang,et al.  Depth Estimation via Affinity Learned with Convolutional Spatial Propagation Network , 2018, ECCV.

[19]  Burr Settles,et al.  Active Learning Literature Survey , 2009 .

[20]  Camillo J. Taylor,et al.  DFuseNet: Deep Fusion of RGB and Sparse Depth Information for Image Guided Dense Depth Completion , 2019, 2019 IEEE Intelligent Transportation Systems Conference (ITSC).

[21]  Xiaogang Wang,et al.  HMS-Net: Hierarchical Multi-Scale Sparsity-Invariant Network for Sparse Depth Completion , 2018, IEEE Transactions on Image Processing.

[22]  Klaus Diepold,et al.  Dense disparity maps from sparse disparity measurements , 2011, 2011 International Conference on Computer Vision.

[23]  Ambuj Tewari,et al.  Active Learning for Non-Parametric Regression Using Purely Random Trees , 2018, NeurIPS.

[24]  David A. Forsyth,et al.  Sparse depth super resolution , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[25]  Andrew McCallum,et al.  Toward Optimal Active Learning through Monte Carlo Estimation of Error Reduction , 2001, ICML 2001.

[26]  Robert D. Nowak,et al.  Faster Rates in Regression via Active Learning , 2005, NIPS.

[27]  Junmo Kim,et al.  Deep Architecture With Cross Guidance Between Single Image and Sparse LiDAR Data for Depth Completion , 2020, IEEE Access.

[28]  M. Pollefeys,et al.  DeepLiDAR: Deep Surface Normal Guided Depth Prediction for Outdoor Scene From Sparse LiDAR Data and Single Color Image , 2018, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[29]  Luc Van Gool,et al.  Sparse and Noisy LiDAR Completion with RGB Guidance and Uncertainty , 2019, 2019 16th International Conference on Machine Vision Applications (MVA).

[30]  Sertac Karaman,et al.  Self-Supervised Sparse-to-Dense: Self-Supervised Depth Completion from LiDAR and Monocular Camera , 2018, 2019 International Conference on Robotics and Automation (ICRA).

[31]  Thomas Brox,et al.  Sparsity Invariant CNNs , 2017, 2017 International Conference on 3D Vision (3DV).

[32]  Jie Tang,et al.  Learning Guided Convolutional Network for Depth Completion , 2019, IEEE Transactions on Image Processing.

[33]  Shai Avidan,et al.  Learning to Sample , 2018, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[34]  Uwe Stilla,et al.  Active learning approach to detecting standing dead trees from ALS point clouds combined with aerial infrared imagery , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[35]  Huikai Xie,et al.  Directionally Controlled Time-of-Flight Ranging for Mobile Sensing Platforms , 2018, Robotics: Science and Systems.

[36]  Steve Hanneke,et al.  Theory of Disagreement-Based Active Learning , 2014, Found. Trends Mach. Learn..

[37]  Hyeran Byun,et al.  Depth map refinement using multiple patch-based depth image completion via local stereo warping , 2010 .

[38]  Reinhard Klette,et al.  Embedded and real-time vehicle detection system for challenging on-road scenes , 2017 .

[39]  Sung Ho Park,et al.  Active semi-supervised learning with multiple complementary information , 2019, Expert Syst. Appl..

[40]  Sertac Karaman,et al.  Sparse-to-Dense: Depth Prediction from Sparse Depth Samples and a Single Image , 2017, 2018 IEEE International Conference on Robotics and Automation (ICRA).

[41]  Hujun Bao,et al.  Depth Completion From Sparse LiDAR Data With Depth-Normal Constraints , 2019, 2019 IEEE/CVF International Conference on Computer Vision (ICCV).

[42]  Ju Shen,et al.  Layer Depth Denoising and Completion for Structured-Light RGB-D Cameras , 2013, 2013 IEEE Conference on Computer Vision and Pattern Recognition.

[43]  Michael Felsberg,et al.  Confidence Propagation through CNNs for Guided Sparse Depth Regression , 2018, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[44]  Raquel Urtasun,et al.  Learning Joint 2D-3D Representations for Depth Completion , 2019, 2019 IEEE/CVF International Conference on Computer Vision (ICCV).

[45]  Gaël Varoquaux,et al.  Scikit-learn: Machine Learning in Python , 2011, J. Mach. Learn. Res..

[46]  Guy Gilboa,et al.  Super-Pixel Sampler: a Data-driven Approach for Depth Sampling and Reconstruction , 2020, 2020 IEEE International Conference on Robotics and Automation (ICRA).

[47]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[48]  Joachim Denzler,et al.  Active Learning for Regression Tasks with Expected Model Output Changes , 2018, BMVC.

[49]  Yehoshua Y. Zeevi,et al.  The farthest point strategy for progressive image sampling , 1997, IEEE Trans. Image Process..

[50]  Camillo J. Taylor,et al.  Depth Completion via Deep Basis Fitting , 2019, 2020 IEEE Winter Conference on Applications of Computer Vision (WACV).

[51]  Pascal Fua,et al.  Learning Active Learning from Data , 2017, NIPS.

[52]  Moncef Gabbouj,et al.  Adaptive sampling for compressed sensing based image compression , 2014, 2014 IEEE International Conference on Multimedia and Expo (ICME).

[53]  Haitao Liu,et al.  A survey of adaptive sampling for global metamodeling in support of simulation-based complex engineering design , 2017, Structural and Multidisciplinary Optimization.

[54]  Naoki Abe,et al.  Query Learning Strategies Using Boosting and Bagging , 1998, ICML.

[55]  Guy Gilboa,et al.  Robust Recovery of Heavily Degraded Depth Measurements , 2016, 2016 Fourth International Conference on 3D Vision (3DV).

[56]  Emanuel Aldea,et al.  Evidential query-by-committee active learning for pedestrian detection in high-density crowds , 2019, Int. J. Approx. Reason..

[57]  Reinhard Radermacher,et al.  Cross-validation based single response adaptive design of experiments for Kriging metamodeling of deterministic computer simulations , 2013 .

[58]  David B. Cole,et al.  Coherent solid-state LIDAR with silicon photonic optical phased arrays. , 2017, Optics letters.

[59]  Feng Liu,et al.  Depth Enhancement via Low-Rank Matrix Completion , 2014, 2014 IEEE Conference on Computer Vision and Pattern Recognition.

[60]  H. Sebastian Seung,et al.  Selective Sampling Using the Query by Committee Algorithm , 1997, Machine Learning.

[61]  Luca Carlone,et al.  Sparse sensing for resource-constrained depth reconstruction , 2016, 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[62]  George C. Runger,et al.  Active Batch Learning with Stochastic Query-by-Forest (SQBF) , 2011 .

[63]  Hengzhu Liu,et al.  Multi-scale features fusion from sparse LiDAR data and single image for depth completion , 2018 .

[64]  Ruigang Yang,et al.  Stereoscopic inpainting: Joint color and depth completion from stereo images , 2008, 2008 IEEE Conference on Computer Vision and Pattern Recognition.

[65]  Pascal Fua,et al.  Geometry in Active Learning for Binary and Multi-class Image Segmentation , 2016, Comput. Vis. Image Underst..

[66]  David A. Cohn,et al.  Minimizing Statistical Bias with Queries , 1996, NIPS.

[67]  Gordon Wetzstein,et al.  Deep Adaptive LiDAR: End-to-end Optimization of Sampling and Depth Completion at Low Sampling Rates , 2020, 2020 IEEE International Conference on Computational Photography (ICCP).