Scholarlink

<p>One of the key problems in 3D object detection is to reduce the accuracy gap between methods based on LiDAR sensors and those based on monocular cameras. A recent monocular-based 3D object detection method using Pseudo-Stereo representation has demonstrated significant capabilities. However, three problems have been discovered in existing practices: (1) relying on a high-performance monocular depth estimator, (2) the generated image suffering from visual holes, deformations, and artifacts, and (3) being difficult to be compatible with geometry-based stereo detectors. In this work, we propose a novel pseudo-stereo 3D detection framework without depth estimation, called PS-SVDM. This framework utilizes a diffusion model to generate a high-quality virtual right view from a left image to mimic the stereo camera signal. With this representation, we can apply various existing stereo image-based detection algorithms. Afterwards, we further explore the application of PS-SVDM in depth-free stereo 3D detection, and the final framework is compatible with most stereo detectors. Experiments conducted on the KITTI-3D Car category show that our method can achieve state-of the-art performance among published monocular 3D detectors. Code will be released.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Depth-free view synthesis from diffusion models for monocular 3D detector in autonomous driving

  • 期刊论文

  • 发表日期：

    2025年5月12日

查看全文
  • Yuguang Shi,
  • Sifan Zhou,
  • Wei Wang,
  • Xiaobo Lu

摘要

One of the key problems in 3D object detection is to reduce the accuracy gap between methods based on LiDAR sensors and those based on monocular cameras. A recent monocular-based 3D object detection method using Pseudo-Stereo representation has demonstrated significant capabilities. However, three problems have been discovered in existing practices: (1) relying on a high-performance monocular depth estimator, (2) the generated image suffering from visual holes, deformations, and artifacts, and (3) being difficult to be compatible with geometry-based stereo detectors. In this work, we propose a novel pseudo-stereo 3D detection framework without depth estimation, called PS-SVDM. This framework utilizes a diffusion model to generate a high-quality virtual right view from a left image to mimic the stereo camera signal. With this representation, we can apply various existing stereo image-based detection algorithms. Afterwards, we further explore the application of PS-SVDM in depth-free stereo 3D detection, and the final framework is compatible with most stereo detectors. Experiments conducted on the KITTI-3D Car category show that our method can achieve state-of the-art performance among published monocular 3D detectors. Code will be released.