Filters








12 Hits in 0.9 sec

ODE-CNN: Omnidirectional Depth Extension Networks [article]

Xinjing Cheng, Peng Wang, Yanqi Zhou, Chenye Guan, Ruigang Yang
2020 arXiv   pre-print
Xinjing Cheng, Chenye Guan are with Robotics and Auto-driving Lab (RAL), Baidu Research, Baidu Inc., {chengxinjing, guanchenye}@baidu.com.  ... 
arXiv:2007.01475v1 fatcat:sz6eicbf6vbbnl6z6stockshdm

IoU Loss for 2D/3D Object Detection [article]

Dingfu Zhou, Jin Fang, Xibin Song, Chenye Guan, Junbo Yin, Yuchao Dai, Ruigang Yang
2019 arXiv   pre-print
In 2D/3D object detection task, Intersection-over-Union (IoU) has been widely employed as an evaluation metric to evaluate the performance of different detectors in the testing stage. However, during the training stage, the common distance loss (, L_1 or L_2) is often adopted as the loss function to minimize the discrepancy between the predicted and ground truth Bounding Box (Bbox). To eliminate the performance gap between training and testing, the IoU loss has been introduced for 2D object
more » ... ction in and . Unfortunately, all these approaches only work for axis-aligned 2D Bboxes, which cannot be applied for more general object detection task with rotated Bboxes. To resolve this issue, we investigate the IoU computation for two rotated Bboxes first and then implement a unified framework, IoU loss layer for both 2D and 3D object detection tasks. By integrating the implemented IoU loss into several state-of-the-art 3D object detectors, consistent improvements have been achieved for both bird-eye-view 2D detection and point cloud 3D detection on the public KITTI benchmark.
arXiv:1908.03851v1 fatcat:jvhspf5acnejlgjngyre4xlsbe

AutoRemover: Automatic Object Removal for Autonomous Driving Videos [article]

Rong Zhang, Wei Li, Peng Wang, Chenye Guan, Jin Fang, Yuhang Song, Jinhui Yu, Baoquan Chen, Weiwei Xu, Ruigang Yang
2019 arXiv   pre-print
Motivated by the need for photo-realistic simulation in autonomous driving, in this paper we present a video inpainting algorithm AutoRemover, designed specifically for generating street-view videos without any moving objects. In our setup we have two challenges: the first is the shadow, shadows are usually unlabeled but tightly coupled with the moving objects. The second is the large ego-motion in the videos. To deal with shadows, we build up an autonomous driving shadow dataset and design a
more » ... ep neural network to detect shadows automatically. To deal with large ego-motion, we take advantage of the multi-source data, in particular the 3D data, in autonomous driving. More specifically, the geometric relationship between frames is incorporated into an inpainting deep neural network to produce high-quality structurally consistent video output. Experiments show that our method outperforms other state-of-the-art (SOTA) object removal algorithms, reducing the RMSE by over 19%.
arXiv:1911.12588v1 fatcat:223s5gmqjbczzcccbb6pt64neq

LiDAR-based Online 3D Video Object Detection with Graph-based Message Passing and Spatiotemporal Transformer Attention [article]

Junbo Yin, Jianbing Shen, Chenye Guan, Dingfu Zhou, Ruigang Yang
2020 arXiv   pre-print
Existing LiDAR-based 3D object detectors usually focus on the single-frame detection, while ignoring the spatiotemporal information in consecutive point cloud frames. In this paper, we propose an end-to-end online 3D video object detector that operates on point cloud sequences. The proposed model comprises a spatial feature encoding component and a spatiotemporal feature aggregation component. In the former component, a novel Pillar Message Passing Network (PMPNet) is proposed to encode each
more » ... crete point cloud frame. It adaptively collects information for a pillar node from its neighbors by iterative message passing, which effectively enlarges the receptive field of the pillar feature. In the latter component, we propose an Attentive Spatiotemporal Transformer GRU (AST-GRU) to aggregate the spatiotemporal information, which enhances the conventional ConvGRU with an attentive memory gating mechanism. AST-GRU contains a Spatial Transformer Attention (STA) module and a Temporal Transformer Attention (TTA) module, which can emphasize the foreground objects and align the dynamic objects, respectively. Experimental results demonstrate that the proposed 3D video object detector achieves state-of-the-art performance on the large-scale nuScenes benchmark.
arXiv:2004.01389v1 fatcat:i46zosrmynabrl66egkeqnrr64

CSPN++: Learning Context and Resource Aware Convolutional Spatial Propagation Networks for Depth Completion [article]

Xinjing Cheng, Peng Wang, Chenye Guan, Ruigang Yang
2019 arXiv   pre-print
Depth Completion deals with the problem of converting a sparse depth map to a dense one, given the corresponding color image. Convolutional spatial propagation network (CSPN) is one of the state-of-the-art (SoTA) methods of depth completion, which recovers structural details of the scene. In this paper, we propose CSPN++, which further improves its effectiveness and efficiency by learning adaptive convolutional kernel sizes and the number of iterations for the propagation, thus the context and
more » ... omputational resources needed at each pixel could be dynamically assigned upon requests. Specifically, we formulate the learning of the two hyper-parameters as an architecture selection problem where various configurations of kernel sizes and numbers of iterations are first defined, and then a set of soft weighting parameters are trained to either properly assemble or select from the pre-defined configurations at each pixel. In our experiments, we find weighted assembling can lead to significant accuracy improvements, which we referred to as "context-aware CSPN", while weighted selection, "resource-aware CSPN" can reduce the computational resource significantly with similar or better accuracy. Besides, the resource needed for CSPN++ can be adjusted w.r.t. the computational budget automatically. Finally, to avoid the side effects of noise or inaccurate sparse depths, we embed a gated network inside CSPN++, which further improves the performance. We demonstrate the effectiveness of CSPN++on the KITTI depth completion benchmark, where it significantly improves over CSPN and other SoTA methods.
arXiv:1911.05377v2 fatcat:dbuq6nvr7jh6fb2dnh5r5af46m

ApolloCar3D: A Large 3D Car Instance Understanding Benchmark for Autonomous Driving [article]

Xibin Song, Peng Wang, Dingfu Zhou, Rui Zhu, Chenye Guan, Yuchao Dai, Hao Su, Hongdong Li, Ruigang Yang
2018 arXiv   pre-print
Autonomous driving has attracted remarkable attention from both industry and academia. An important task is to estimate 3D properties(e.g.translation, rotation and shape) of a moving or parked vehicle on the road. This task, while critical, is still under-researched in the computer vision community - partially owing to the lack of large scale and fully-annotated 3D car database suitable for autonomous driving research. In this paper, we contribute the first large-scale database suitable for 3D
more » ... ar instance understanding - ApolloCar3D. The dataset contains 5,277 driving images and over 60K car instances, where each car is fitted with an industry-grade 3D CAD model with absolute model size and semantically labelled keypoints. This dataset is above 20 times larger than PASCAL3D+ and KITTI, the current state-of-the-art. To enable efficient labelling in 3D, we build a pipeline by considering 2D-3D keypoint correspondences for a single instance and 3D relationship among multiple instances. Equipped with such dataset, we build various baseline algorithms with the state-of-the-art deep convolutional neural networks. Specifically, we first segment each car with a pre-trained Mask R-CNN, and then regress towards its 3D pose and shape based on a deformable 3D car model with or without using semantic keypoints. We show that using keypoints significantly improves fitting performance. Finally, we develop a new 3D metric jointly considering 3D pose and 3D shape, allowing for comprehensive evaluation and ablation study. By comparing with human performance we suggest several future directions for further improvements.
arXiv:1811.12222v2 fatcat:3eu7nofbevbenkpdxddufya7w4

Association of single nucleotide polymorphisms in the RAB5B gene 3'UTR region with polycystic ovary syndrome in Chinese Han women

Jia Yu, Caifei Ding, Siqi Guan, Chenye Wang
2019 Bioscience Reports  
Objective: Previous genome-wide sequencing revealed that Ras-related protein Rab-5B (RAB5B) is a susceptible target in patients with polycystic ovary syndrome (PCOS).Methods: Direct sequencing was performed to analyze the RAB5B gene rs1045435, rs11550558, rs34962186, rs705700, rs58717357, rs11171718, rs60028217, rs772920 loci genotypes in 300 PCOS patients and 300 healthy controls. The plasma microRNA (miRNA)-24, miR-320 levels were measured by reverse transcription fluorescent quantitative PCR
more » ... (RT-qPCR).Results: The risk of PCOS in C allele carriers of RAB5B gene rs1045435 locus was 3.91 times higher than that of G allele. The risk of PCOS in rs11550558 locus G allele was 4.09 times higher than A allele. The risk of PCOS in rs705700 locus C allele was 1.66 times greater than T allele. The risk of PCOS in rs11171718 locus A allele carrier was 3.84 times higher than G allele. The rs11550558 SNP was associated with PCOS risk only in those with age ≥ 31.1 years. And RAB5B gene rs11550558, rs1045435, and rs11171718 SNPs were significantly associated with PCOS risk only in subjects with BMI ≥ 23.8 kg/m2 We also found that the RAB5B gene rs1045435 SNP was associated with plasma miR-24 levels. The RAB5B gene rs11550558, rs705700, rs11171718 SNPs were correlated with plasma miR-230 levels.Conclusion: The single nucleotide polymorphisms of the rs1045435, rs11550558, rs705700, and rs11171718 loci of the RAB5B gene are associated with PCOS risk. The rs1045435 locus is likely an miR-24 binding site, while rs11550558, rs705700, and rs11171718 loci may be miR-320 binding sites.
doi:10.1042/bsr20190292 pmid:31036605 pmcid:PMC6522744 fatcat:boxkn5jbuzfchbkple56cevseq

CSPN++: Learning Context and Resource Aware Convolutional Spatial Propagation Networks for Depth Completion

Xinjing Cheng, Peng Wang, Chenye Guan, Ruigang Yang
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
Depth Completion deals with the problem of converting a sparse depth map to a dense one, given the corresponding color image. Convolutional spatial propagation network (CSPN) is one of the state-of-the-art (SoTA) methods of depth completion, which recovers structural details of the scene. In this paper, we propose CSPN++, which further improves its effectiveness and efficiency by learning adaptive convolutional kernel sizes and the number of iterations for the propagation, thus the context and
more » ... omputational resource needed at each pixel could be dynamically assigned upon requests. Specifically, we formulate the learning of the two hyper-parameters as an architecture selection problem where various configurations of kernel sizes and numbers of iterations are first defined, and then a set of soft weighting parameters are trained to either properly assemble or select from the pre-defined configurations at each pixel. In our experiments, we find weighted assembling can lead to significant accuracy improvements, which we referred to as "context-aware CSPN", while weighted selection, "resource-aware CSPN" can reduce the computational resource significantly with similar or better accuracy. Besides, the resource needed for CSPN++ can be adjusted w.r.t. the computational budget automatically. Finally, to avoid the side effects of noise or inaccurate sparse depths, we embed a gated network inside CSPN++, which further improves the performance. We demonstrate the effectiveness of CSPN++ on the KITTI depth completion benchmark, where it significantly improves over CSPN and other SoTA methods 1.
doi:10.1609/aaai.v34i07.6635 fatcat:rt5jzhgjzvd2hiu74zf4axeumy

AutoRemover: Automatic Object Removal for Autonomous Driving Videos

Rong Zhang, Wei Li, Peng Wang, Chenye Guan, Jin Fang, Yuhang Song, Jinhui Yu, Baoquan Chen, Weiwei Xu, Ruigang Yang
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
Motivated by the need for photo-realistic simulation in autonomous driving, in this paper we present a video inpainting algorithm AutoRemover, designed specifically for generating street-view videos without any moving objects. In our setup we have two challenges: the first is the shadow, shadows are usually unlabeled but tightly coupled with the moving objects. The second is the large ego-motion in the videos. To deal with shadows, we build up an autonomous driving shadow dataset and design a
more » ... ep neural network to detect shadows automatically. To deal with large ego-motion, we take advantage of the multi-source data, in particular the 3D data, in autonomous driving. More specifically, the geometric relationship between frames is incorporated into an inpainting deep neural network to produce high-quality structurally consistent video output. Experiments show that our method outperforms other state-of-the-art (SOTA) object removal algorithms, reducing the RMSE by over 19%.
doi:10.1609/aaai.v34i07.6982 fatcat:kgrz7qzw6zhmrcsdoearrzzwna

LIGA-Stereo: Learning LiDAR Geometry Aware Representations for Stereo-based 3D Detector [article]

Xiaoyang Guo, Shaoshuai Shi, Xiaogang Wang, Hongsheng Li
2021 arXiv   pre-print
tional Conference on Robotics and Automation (ICRA), [70] Dingfu Zhou, Jin Fang, Xibin Song, Chenye Guan, Junbo pages 2379–2384. IEEE, 2019.  ...  In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, pages 12926–12934, 2020. 3 [10] Dingfu Zhou, Jin Fang, Xibin Song, Chenye Guan, Junbo Yin, Yuchao Dai, and Ruigang  ... 
arXiv:2108.08258v1 fatcat:hz7ix7azmfbf3oli7mpay7jf3a

Table of Contents

2021 2021 40th Chinese Control Conference (CCC)   unpublished
XU Mingkuan, YU Yang, WU Chenye 1055 Linear Quadratic Stackelberg Stochastic Differential Games: Closed-Loop Solvability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ...  LEI Zhongcheng, ZHOU Hong, HU Wenshan, LIU Guo-Ping, GUAN Shiqi, FENG Xingle 8715 Surface Texture Recognition Network Based on Flexible Electronic Skin . . . . . . . . . . . .  ... 
doi:10.23919/ccc52363.2021.9550117 fatcat:55y7a2gagfhtpc6llmfvl7gqpm

An Embarrassingly Pragmatic Introduction to Vision-based Autonomous Robots [article]

Marcos V. Conde
2021 arXiv   pre-print
[79] Xibin Song, Peng Wang, Dingfu Zhou, Rui Zhu, Chenye Guan, Yuchao Dai, Hao Su, Hongdong Li, and Ruigang Yang.  ... 
arXiv:2112.05534v2 fatcat:3drhsxelvvdwvpsq5rvfpnukam