We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CV

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computer Vision and Pattern Recognition

Title: PICCOLO: Point Cloud-Centric Omnidirectional Localization

Abstract: We present PICCOLO, a simple and efficient algorithm for omnidirectional localization. Given a colored point cloud and a 360 panorama image of a scene, our objective is to recover the camera pose at which the panorama image is taken. Our pipeline works in an off-the-shelf manner with a single image given as a query and does not require any training of neural networks or collecting ground-truth poses of images. Instead, we match each point cloud color to the holistic view of the panorama image with gradient-descent optimization to find the camera pose. Our loss function, called sampling loss, is point cloud-centric, evaluated at the projected location of every point in the point cloud. In contrast, conventional photometric loss is image-centric, comparing colors at each pixel location. With a simple change in the compared entities, sampling loss effectively overcomes the severe visual distortion of omnidirectional images, and enjoys the global context of the 360 view to handle challenging scenarios for visual localization. PICCOLO outperforms existing omnidirectional localization algorithms in both accuracy and stability when evaluated in various environments.
Comments: Accepted to ICCV 2021
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:2108.06545 [cs.CV]
  (or arXiv:2108.06545v2 [cs.CV] for this version)

Submission history

From: Junho Kim [view email]
[v1] Sat, 14 Aug 2021 14:19:37 GMT (7611kb,D)
[v2] Thu, 14 Oct 2021 23:54:50 GMT (7611kb,D)

Link back to: arXiv, form interface, contact.