Balancing accuracy and efficiency in lumbar spine image segmentation using multi-scale attention and residual learning
Vijaya Bhaskar Sadu, Bharathi Subramaniam, Rukmani Devi Sethuraman, Sudhakar Tummala, Beluguri Venkateswarlu, Jeevika Tharini Viswanathan · 2026
Read Balancing accuracy and efficiency in lumbar spine image segmentation using multi-scale attention and residual learning by Vijaya Bhaskar Sadu, Bharathi Subramaniam, Rukmani Devi Sethuraman, Sudhakar Tummala, Beluguri Venkateswarlu, Jeevika Tharini Viswanathan (2026) 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/PMC13538668.txt.
Description
Vijaya Bhaskar Sadu and colleagues design a deep learning model for outlining lower spine structures in MRI scans in a 2026 study. Accurate outlining of vertebrae and discs guides diagnosis and surgery, yet precision usually costs heavy computing. The network pairs multi-scale attention blocks with adaptive feature selection across pyramid levels and gated links between encoder and decoder. Training used 215 expert-labeled lumbar scan volumes.
Overlap scores beat three established rival models by clear margins, and boundary error measures improved by roughly a fifth to a quarter over the best baseline. Quality held steady across all spine levels and discs, with ablations crediting each new block. Training time and memory stayed manageable. The authors position the model for surgical planning and diagnostic support.
Catalog record
| CREATOR | Vijaya Bhaskar Sadu, Bharathi Subramaniam, Rukmani Devi Sethuraman, Sudhakar Tummala, Beluguri Venkateswarlu, Jeevika Tharini Viswanathan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 116292 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |