Neighbor-Vote

Chu, Xiaomeng; Deng, Jiajun; Yao, Li; Yuan, Zhenxun; Zhang, Yanyong; Ji, Jianmin

doi:10.1145/3474085.3475641

Cited by 16 publications

(2 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The architectures for monocular 3D object detection can be mainly categorized into two groups: Pseudo-LiDAR based methods [11,34,43] that use pre-trained depth networks to generate pseudo LiDAR representations, e.g. pseudo point clouds and pseudo voxels, and then feed them to LiDAR-based 3D detectors, and the rest monocular 3D detection methods that use 2D feature learning from a single image with optional 3D cues matching, concatenating or guiding for 3D perception [24,26,31,38,39,54,57].…”

Section: Related Workmentioning

confidence: 99%

“…However, the huge potential in such a cheap and easy-to-deploy solution to 3D detection attracts more and more researchers. Remarkable progress has been made in Pseudo-LiDAR detectors [11,29,34,43,52] that use a pre-trained depth estimation network to generate Pseudo-LiDAR representations, e.g. pseudo point clouds and pseudo voxels, and then feed them to LiDAR-based 3D detectors.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Pseudo-Stereo for Monocular 3D Object Detection in Autonomous Driving

Chen

Dai

Ding

2022

2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

View full text Add to dashboard Cite

Pseudo-LiDAR 3D detectors have made remarkable progress in monocular 3D detection by enhancing the capability of perceiving depth with depth estimation networks, and using LiDAR-based 3D detection architectures. The advanced stereo 3D detectors can also accurately localize 3D objects. The gap in image-to-image generation for stereo views is much smaller than that in image-to-LiDAR generation. Motivated by this, we propose a Pseudo-Stereo 3D detection framework with three novel virtual view generation methods, including image-level generation, featurelevel generation, and feature-clone, for detecting 3D objects from a single image. Our analysis of depth-aware learning shows that the depth loss is effective in only feature-level virtual view generation and the estimated depth map is effective in both image-level and feature-level in our framework. We propose a disparity-wise dynamic convolution with dynamic kernels sampled from the disparity feature map to filter the features adaptively from a single image for generating virtual image features, which eases the feature degradation caused by the depth estimation errors. Till submission (November 18, 2021), our Pseudo-Stereo 3D detection framework ranks 1 st on car, pedestrian, and cyclist among the monocular 3D detectors with publications on the KITTI-3D benchmark. The code is released at https: //github.com/revisitq/Pseudo-Stereo-3D.

show abstract