TSE-Net: Semi-supervised Monocular Height Estimation from Single Remote Sensing Images Journal: ISPRS Journal of Photogrammetry and Remote Sensing
Sining Chen, Xiao Xiang Zhu · 2025
Read TSE-Net: Semi-supervised Monocular Height Estimation from Single Remote Sensing Images Journal: ISPRS Journal of Photogrammetry and Remote Sensing by Sining Chen, Xiao Xiang Zhu (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-2511-13552.txt.
Description
Sining Chen and Xiao Xiang Zhu present TSE-Net, a semi-supervised method that estimates ground height from a single remote sensing image. Height from one image is a cheap stand-in for multi-view or LiDAR mapping, but labeled examples are costly and scarce. Their pipeline learns from large pools of unlabeled images instead of waiting for more labels.
A teacher network creates pseudo-labels, a student learns from them, and an exam network averages the student over time to steady results. The teacher also classifies heights into classes built by a hierarchical split and filters weak labels by calibrated confidence. Tests on three datasets with varied resolution and sensors showed the approach holds up across settings.
Catalog record
| CREATOR | Sining Chen, Xiao Xiang Zhu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 54 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-13552