Object scene flow for autonomous vehicles

By Moritz Menze, Andreas Geiger. 2015

2,263 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition, 3D Vision & Pose, Robotics & Control

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

In 3D Vision & Pose

In Robotics & Control

Citations in the corpus

Perceiver IO: A General Architecture for Structured Inputs & Outputs (2021, neutral): “Network Sintel.clean Sintel.final KITTI PWCNet [75] 2.17 2.91 5.76 RAFT [84] 1.95 2.57 4.23 Perceiver IO 1.81 2.42 4.98 to find correspondences.”
Scene Flow to Action Map: A New Representation for RGB-D Based Action Recognition with Convolutional Neural Networks (2017, background): “In contrast to optical flow which provides the projection of the scene motion onto the image plane, scene flow [41, 6, 26, 8, 11, 37, 32] estimates the actual 3D motion field.”
A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow Estimation (2015, methodology): “The KITTI dataset was produced in 2012 [8] and extended in 2015 [16].”
Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-Shot Cross-Dataset Transfer (2019, methodology): “Commonly used datasets feature homogeneous scene layouts, such as street scenes in a specific geographic region [3], [28], [29] or indoor environments [30].”
Playing for Benchmarks (2017, background): “For example, optical flow ground truth is sparse and is based on fitting approximate CAD models to moving objects [42], object detection and segmentation data is available for only a small number of frames and a small set of classes [63], and the dataset as a whole represents the appearance of a sin”
Spatio-temporal video autoencoder with differentiable memory (2015, background): “The dense ﬂow map is useful to explain dynamic environments, where different objects can follow different motion models, hence a global transformation is not sufﬁcient Menze & Geiger (2015).”
Stereo Matching by Training a Convolutional Neural Network to Compare Image Patches (2015, methodology): “The KITTI stereo data set (Geiger et al., 2013; Menze and Geiger, 2015) is a collection of rectiﬁed image pairs taken from two video cameras mounted on the roof of a car, roughly 54 centimeters apart.”
Learning 3D Shape Completion from Laser Scan Data with Weak Supervision (2018, neutral): “, [31,50]).”
SelFlow: Self-Supervised Learning of Optical Flow (2019, methodology): “We evaluate and compare our methods with stateof-the-art unsupervised and supervised learning methods on public optical flow benchmarks including MPI Sintel [7], KITTI 2012 [11] and KITTI 2015 [30].”
UnFlow: Unsupervised Learning of Optical Flow with a Bidirectional Census Loss (2017, methodology): “For the KITTI benchmark, FlowNet2 is fine-tuned on the sparse ground truth from the KITTI 2012 (Geiger, Lenz, and Urtasun 2012) and KITTI 2015 (Menze and Geiger 2015) training sets and achieves state-of-the-art accuracy.”
Hierarchical Neural Architecture Search for Deep Stereo Matching (2020, methodology): “We then conduct the architecture evaluation on KITTI 2012 [29], KITTI 2015 [30] and Middlebury 2014 [31] benchmarks by inheriting the searched architecture from SceneFlow dataset.”
A Survey on Deep Learning Techniques for Stereo-Based Depth Estimation (2020, background): “474 Middleburry [20], KITTI2015 [21], 9.”
ESPNet: Efficient Spatial Pyramid of Dilated Convolutions for Semantic Segmentation (2018, background): “This is especially problematic on the computationally heavy task of semantic segmentation [4,5,6,7,8,9,10].”
SfM-Net: Learning of Structure and Motion from Video (2017, methodology): “We show results of our approach on KITTI [12, 21], MoSeg [4], and RGB-D SLAM [27] benchmarks under different levels of supervision.”
A survey on face data augmentation for the training of deep neural networks (2019, background): “More experiences of data augmentation by AR can be found in the studies of [3, 31, 100].”
Real-Time Self-Adaptive Deep Stereo (2018, background): “As mentioned above, the Table also reports the performance of the models fine-tuned offline on the 400 stereo pairs with groundtruth disparities from the KITTI 2012 and 2015 training dataset [23, 8].”
DDFlow: Learning Optical Flow with Unlabeled Data Distillation (2019, background): “Optical flow estimation is a core computer vision building block, with a wide range of applications, including autonomous driving (Menze and Geiger 2015), object tracking (Chauhan and Krishan 2013), action recognition (Simonyan and Zisserman 2014) and video processing (Bonneel et al.”
DSNet: Joint Learning for Scene Segmentation and Disparity Estimation (2019, methodology): “For our disparity estimation on the KITTI 2015 [27], we mainly pretrain the model on CityScapes dataset and then fine tune on KITTI Stereo dataset, using 75% of the train set of KITTI as our train set while testing on the remaining samples.”