Joint SFM and detection cues for monocular 3D localization in road scenes

By Shiyu Song, Manmohan Chandraker. 2015

87 citations recorded by the source corpus.

Subfields: 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 3D Vision & Pose

1. Yin Zhou
2. Oncel Tuzel
3. Bo Li
4. Vishwanath A. Sindagi
5. Raquel Urtasun
6. Xiaozhi Chen
7. Binh Yang
8. Xiang Bai

In Robotics & Control

1. Yin Zhou
2. Bo Li
3. Oncel Tuzel
4. Shichao Yang
5. Sebastian A. Scherer
6. Vishwanath A. Sindagi
7. Manmohan Chandraker
8. Huimin Ma

Citations in the corpus

• VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection (2017, background): “Given that images provide detailed texture information, many algorithms infered the 3D bounding boxes from 2D images [4, 3, 44, 45, 46, 38].”
• Multi-view 3D Object Detection Network for Autonomous Driving (2016, methodology): “e with 3DOP, it generates 3D proposals from monocular images. [31,32] introduces a detailed geometry representation of objects using 3D wireframe models. To incorporate temporal information, some work[6,21] combine structure from motion and ground estimation to lift 2D detection boxes to 3D bounding”
• TANet: Robust 3D Object Detection from Point Clouds with Triple Attention (2019, background): “…has been devoted to detect 3D objects from multi-view images, which can be roughly classified into two categories: the front view based approaches (Song and Chandraker 2015; Chen et al. 2016; Mousavian et al. 2017), and the bird’s eye view based approaches (Chen et al. 2017; Ku et al. 2018;…”
• MVX-Net: Multimodal VoxelNet for 3D Object Detection (2019, methodology): “Different from these approaches, image-based methods ([3], [4], [39], [44], [45], [34], [2]) were developed to infer 3D bounding boxes from 2D images.”
• CubeSLAM: Monocular 3-D Object SLAM (2018, methodology): “bject Localization Comparison on KITTI Raw Sequence Seq 04 47 56 Obj. ID 1 2 3 6 0 4 9 12 0 Mean No. Frames 91 251 284 169 170 96 94 637 293 Depth [51] 4.1 6.8 5.3 7.3 9.6 11.4 7.1 10.5 5.5 7.9 error [32] 6.0 5.6 4.9 5.9 5.9 12.5 7.0 8.2 6.0 6.8 (%) Ours 5.1 11.1 1.8 2.5 9.3 3.8 6.6 2.8 0.8 4.9 Time”
• A Continuous Occlusion Model for Road Scene Understanding (2016, background): “Prior works such as [24] combine information from point tracks and detection bounding boxes, but do not consider occlusions for either.”