Are Cars Just 3D Boxes? Jointly Estimating the 3D Shape of Multiple Objects

By M. Zia, Michael Stark, Konrad Schindler. 2014

83 citations recorded by the source corpus.

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

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

1. Xiaogang Wang
2. Shaoshuai Shi
3. Hongsheng Li
4. Yin Zhou
5. Jianping Shi
6. Bo Li
7. Li Jiang
8. Oncel Tuzel

In Computer Vision and Pattern Recognition

1. Luc Van Gool
2. Xiaogang Wang
3. Ming-Hsuan Yang
4. Dacheng Tao
5. Jitendra Malik
6. Trevor Darrell
7. Alan L. Yuille
8. Bernt Schiele

In Geology

1. Li Jiang
2. Shaoshuai Shi
3. Xiaogang Wang
4. Hongsheng Li
5. Zhe Wang
6. Jianping Shi
7. Jiajun Deng
8. Chaoxu Guo

In 3D Vision & Pose

1. Yin Zhou
2. Oncel Tuzel
3. Shaoshuai Shi
4. Xiaogang Wang
5. Hongsheng Li
6. Bo Li
7. Jianping Shi
8. Raquel Urtasun

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): “[27, 28] introduces a detailed geometry representation of objects using 3D wireframe models.”
• Joint SFM and detection cues for monocular 3D localization in road scenes (2015, background): “Recent scene understanding frameworks also reason about object relationships to handle occlusions in monocular input [23].”
• 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.”
• A General Pipeline for 3D Detection of Vehicles (2018, methodology): “In [12], [30], [11] and [7], 3D wireframe models were used.”
• PV-RCNN++: Point-Voxel Feature Set Abstraction With Local Vector Representation for 3D Object Detection (2021, background): “[45], [46], [47], [48], [49] develop monocular 3D object detectors based on a wire-frame template obtained from CAD models.”
• Stereo R-CNN Based 3D Object Detection for Autonomous Driving (2019, background): “[30, 1, 22] explicitly utilize sparse information by predicting series of keypoints of regular-shape vehicles.”
• 3D-RCNN: Instance-Level 3D Object Reconstruction via Render-and-Compare (2018, methodology): “When it comes to shape, most methods either estimate only 3D bounding boxes [36, 6, 12], or coarse wire-frame skeletons [29, 54, 63, 62], or represent shape via an exemplar mesh chosen from a small set of meshes [4, 56, 35, 2].”
• A large-scale car dataset for fine-grained categorization and verification (2015, background): “Other car-related research includes detection [22], tracking [17] [24], joint detection and pose estimation [6, 25], and 3D parsing [30].”
• Data-driven 3D Voxel Patterns for object category recognition (2015, background): “We note that [44] also evaluates 3D localization on KITTI images.”
• A Continuous Occlusion Model for Road Scene Understanding (2016, background): “This is in contrast to prior works that consider occlusions in 2D, or through discrete occluder patterns or models that are not physically interpretable [18, 19, 30, 33, 34, 35].”
• Disentangling Monocular 3D Object Detection (2019, background): “In [35], the recognition task is tackled by jointly reasoning”
• Learning Depth-Guided Convolutions for Monocular 3D Object Detection (2019, methodology): “To introduce more prior information, [2, 24, 57, 53] used 3D shapes as templates to get better object geometry.”
• Orthographic Feature Transform for Monocular 3D Object Detection (2018, methodology): “Many approaches start from 2D bounding boxes extracted using standard detectors described above, upon which they either directly regress 3D pose parameters for each region [14, 26, 24, 23] or fit 3D templates to the image [2, 35, 36, 38].”