NextStop: An Improved Tracker For Panoptic LIDAR Segmentation Data
Nirit Alkalay, Roy Orfaig, Ben-Zion Bobrovsky · 2025
Read NextStop: An Improved Tracker For Panoptic LIDAR Segmentation Data by Nirit Alkalay, Roy Orfaig, Ben-Zion Bobrovsky (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-06235.txt.
Description
Self-driving systems must label every point in a laser scan and keep each object's identity steady over time. Nirit Alkalay, Roy Orfaig, and Ben-Zion Bobrovsky presented a tracker for that task in 2025, called NextStop. Earlier trackers relied on short-term detection only, ignored motion estimates, and dropped small objects, which caused frequent identity switches between frames.
NextStop adds motion estimation with Kalman filters, careful matching of objects across frames, and management of how long a track should live. A tracklet state system ranks which tracks to trust. On a standard validation set, it tracked small objects such as people and cyclists better, started tracks earlier, and switched identities less often.
Catalog record
| CREATOR | Nirit Alkalay, Roy Orfaig, Ben-Zion Bobrovsky |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 52 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-06235