PDSE: A Multiple Lesion Detector for CT Images using PANet and Deformable Squeeze-and-Excitation Block
Di Fan, Heng Yu, Zhiyuan Xu · 2025
Read PDSE: A Multiple Lesion Detector for CT Images using PANet and Deformable Squeeze-and-Excitation Block by Di Fan, Heng Yu, Zhiyuan Xu (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-2506-03608.txt.
Description
Di Fan, Heng Yu, and Zhiyuan Xu present a lesion detector for CT scans in a 2025 paper. Finding lesions in these scans is hard because they vary in type, size, and place. The authors build a one-stage detector, called PDSE, by reworking a known network named RetinaNet.
They improve how features flow through the network by adding a low-level feature map to the path aggregation step. They also add a flexible attention block that helps the model weigh channels of information. On the public DeepLesion benchmark, the detector set a new best score. The gains were strongest for small lesions and lesions at mixed sizes, which are often the hardest to catch.
Catalog record
| CREATOR | Di Fan, Heng Yu, Zhiyuan Xu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 15 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-03608